A Study Utilizing Small Segment Matching

There has been quite a bit of discussion in the last several weeks, both pro and con, about how to use small matching DNA segments in genetic genealogy.  A couple of people are even of the opinion that small segments can’t be used at all, ever.  Others are less certain and many of us are working our way through various scenarios.  Evidence certainly exists that these segments can be utilized.

I’ve been writing foundation articles, in preparation for this article, for several weeks now.  Recently, I wrote about how phasing works and determining IBD versus IBS matches and included guidelines for telling the difference between the different kinds of matches.  If you haven’t read that article, it’s essential to understanding this article, so now would be a good time to read or review that article.

I followed that with a step by step article, Demystifying Autosomal DNA Matching, on how to do phasing and matching in combination with the guidelines about how to determine IBD (identical by descent) versus IBS (identical by chance) and identical by population matches when evaluating your own matches.

Now that we understand IBS, IBD, Phasing and how matching actually works on a case by case basis, let’s look at applying those same matching and IBS vs IBD guidelines to small data segments as well.

A Little History

So those of you who haven’t been following the discussion on various blogs and social media don’t feel like you’ve been dropped into the middle of a conversation with no context, let me catch you up.

On Thanksgiving Day, I published an article about identifying one of my ancestors, after many years of trying, Sarah Hickerson.

That article spurred debate, which is just fine when the debate is about the science, but it subsequently devolved into something less pleasant.  There are some individuals with very strong opinions that utilizing small segments of DNA data can “never be done.”

I do not agree with that position.  In fact, I strongly disagree and there are multiple cases with evidence to support small segments being both accurate and useful in specific types of genealogical situations.  We’ll take a look at several.

I do agree that looking at small segment data out of context is useless.  To the best of my knowledge, no genealogist begins with their smallest segments and tries to assemble them, working from the bottom up.  We all begin with the largest segments, because they are the most useful and the closest connections in our tree, and work our way down.  Generally, we only work with small segments when we have to – and there are times that’s all we have.  So we need to establish guidelines and ways to know if those small segments are reliable or not.  In other words, how can we draw conclusions and how much confidence can we put in those conclusions?

Ultimately, whether you choose to use or work with small segment data will be your own decision, based on your own circumstances.  I simply wanted to understand what is possible and what is reasonable, both for my own genealogy and for my readers.

In my projects, I haven’t been using small segment data out of context, or randomly.  In other words, I don’t just pick any two small segment matches and infer or decide that they are valid matches.  Fortunately, by utilizing the IBD vs IBS guidelines, we have tools to differentiate IBD (Identical by Descent) segments from IBS (Identical by State) by chance segments and IBD/IBS by population for matching segments, both large and small.

Studying small segment data is the key to determining exactly how small segments can reasonably be utilized.  This topic probably isn’t black or white, but shades of gray – and assuming the position that something can’t be done simply assures that it won’t be.

I would strongly encourage those involved and interested in this type of research to retain those small segments, work with them and begin to look for patterns.  The only way we, as a community, are ever going to figure out how to work with small segments successfully and reliably is to, well, work with them.

Discussing the science and scenarios surrounding the usage of small data segments in various different situations is critical to seeing our way through the forest.  If the answers were cast in concrete about how to do this, we wouldn’t be working through this publicly today.

Negative personal comments and inferences have no place in the scientific community.  It discourages others from participating, and serves to stifle research and cooperation, not encourage it.  I hope that civil scientific discussions and comparisons involving small segment data can move forward, with decorum, because they are critically needed in order to enhance our understanding, under varying circumstances, of how to utilize small segment data.  As Judy Russell said, disagreeing doesn’t have to be disagreeable.

Two bloggers, Blaine Bettinger and CeCe Moore wrote articles following my Hickerson article.  Blaine subsequently wrote a second article here.  Felix Immanuel wrote articles here and here.

A few others have weighed in, in writing, as well although most commentary has been on Facebook.  Israel Pickholtz, a professional genealogist and genetic consultant, stated on his blog, All My Foreparents, the following:

It is my nature to distrust rules that put everything into a single category and that’s how I feel about small segments. Sometimes they are meaningful and useful, sometimes not.

When I reconstructed my father’s DNA using Lazerus (described last week in Genes From My Father), I happily accepted all small segments of whatever size because those small segments were in the DNA of at least one of his children and at least one of his brother/sister/first cousin. If I have a particular small segment, I must have received it from my parents. If my father’s brother (or sister) has it as well, then it is eminently clear to me that I got it from my father and that it came to him and his brother from my grandfather. And it is not reasonable to say that a sliver of that small segment might have come from my mother, because my father’s people share it.

After seeing Israel’s commentary about Lazarus, I reconstructed the genome of both Roscoe and John Ferverda, brothers, which includes both large and small segments.  Working with the Ferverda DNA further, I wrote an article, Just One Cousin, about matching between two siblings and a first cousin, which includes lots of small data segments, some of which were proven to triangulate, meaning they are genuine, and some which did not.  There are lots more examples in the demystifying article, as well.

What Not To Do 

Before we begin, I want to make it very clear that am not now, and never have, advocated that people utilize small data segments out of context of larger matching segments and/or at least suspected matching genealogy.  For example, I have never implied or even hinted that anyone should go to GedMatch, do a “one to many” compare at 1 cM and then contact people informing them that they are related.  Anyone who has extrapolated what I’ve written to mean that either simply did not understand or intentionally misinterpreted the articles.

Sarah Hickerson Revisited

If I thought Sarah Hickerson caused me a lot of heartburn in the decades before I found her, little did I know how much heartburn that discovery would cause.

Let’s go back to the Sarah Hickerson article that started the uproar over whether small data segments are useful at all.

In that article, I found I was a member of a new Ancestry DNA Circle for Charles Hickerson and Mary Lytle, the parents of Sarah Hickerson.

Ancestry Hickerson match

Because there are no tools at Ancestry to prove DNA connections, I hurried over to Family Tree DNA looking for any matches to Hickersons for myself and for my Vannoy cousins who also (potentially) descended from this couple.  Much to my delight, I found  several matches to Hickersons, in fact, more than 20 – a total of 614 rows of spreadsheet matches when I included all of my Vannoy cousins who potentially descend from this couple to their Hickerson matches.  There were 64 matching clusters of segments, both small and large.  Some matches were as large as 20cM with 6000 SNPs and more than 20 were over 10cM with from 1500 to 6000 SNPs.  There were also hundreds of small segments that matched (and triangulated) as well.

By the time I added in a few more Vannoy cousins that we’ve since recruited, the spreadsheet is now up to 1093 rows and we have 52 Vannoy-Hickerson TRIANGULATED CLUSTERS utilizing only Family Tree DNA tools.

Triangulated DNA, found in 3 or more people at the same location who share a common ancestor is proven to be from that ancestor (or ancestral couple.)  This is the commonly accepted gold standard of autosomal DNA triangulation within the industry.

Here’s just one example of a cluster of three people.  Charlene and Buster are known (proven, triangulated) cousins and Barbara is a descendant of Charles Hickerson and Mary Lytle.

example triang

What more could you want?

Yes, I called this a match.  As far as I’m concerned, it’s a confirmed ancestor.  How much more confirmed can you get?

Some clusters have as many as 25 confirmed triangulated members.

chr 13 group

Others took issue with this conclusion because it included small segment data.  This seems like the perfect opportunity in which to take a look at how small segments do, or don’t stand up to scrutiny.  So, let’s do just that.  I also did the same type of matching comparison in a situation with 2 siblings and a known cousin, here.

To Trash…or Not To Trash

Some genetic genealogists discard small segments entirely, generally under either 5 or 7cM, which I find unfortunate for several reasons.

  1. If a person doesn’t work with small segments, they really can’t comment on the lack of results, and they’ll never have a success because the small segments will have been discarded.
  2. If a person doesn’t work with small segments, they will never notice any trends or matches that may have implications for their ancestry.
  3. If a person doesn’t work with small segments, they can’t contribute to the body of evidence for how to reasonably utilize these segments.
  4. If a person doesn’t work with small segments, they may well be throwing the baby out with the bathwater, but they’ll never know.
  5. They encourage others to do the same.

The Sarah Hickerson article was not meant as a proof article for anything – it was meant to be an article encouraging people to utilize genetic genealogy for not only finding their ancestor and proving known connections, but breaking down brick walls.  It was pointing the way to how I found Sarah Hickerson.  It was one of my 52 Ancestors Series, documenting my ancestors, not one of the specifically educational articles.  This article is different.

If you are only interested in the low hanging fruit, meaning within the past 5 or 6 generations, and only proving your known pedigree, not finding new ancestors beyond that 5-6 generation level, then you can just stop reading now – and you can throw away your small segments.  But if you want more, then keep reading, because we as a community need to work with small segment data in order to establish guidelines that work relative to utilizing small segments and identifying the small segments that can be useful, versus the ones that aren’t.

I do not believe for one minute that small segments are universally useless.  As Israel said, if his family did not receive those segments from a common family member, then where did they all get those matching segments?

In fact, utilizing triangulated and proven DNA relationships within families is how adoptees piece together their family trees, piggybacking off of the work of people with known pedigrees that they match genetically.  My assumption had been that the adoptee community utilized only large DNA segments, because the larger the matching segments, generally the closer in time the genealogy match – and theoretically the easier to find.

However, I discovered that I was wrong, and the adoptee community does in fact utilize small segments as well.  Here’s one of the comments posted on my Chromosome Browser War blog article.

“Thanks for the well thought out article, Roberta, I have something to add from the folks at DNAadoption. Adoptees are not just interested in the large segments, the small segments also build the proof of the numerous lines involved. In addition, the accumulation of surnames from all the matches provides a way to evaluate new lines that join into the tree.”

Diane Harman-Hoog (on behalf of the 6 million adoptees in this country, many of who are looking for information on medical records and family heritage).

Diane isn’t the only person who is working with small segment data.  Tim Janzen works with small segments, in particular on his Mennonite project, and discusses small segments on the ISOGG WIKI Phasing page.  Here is what Tim has to say:

“One advantage of Family Finder is that FF has a 1 cM threshold for matching segments. If a parent and a child both have a matching segment that is in the 2 to 5 cM range and if the number of matching SNPs is 500 or more then there is a reasonably high likelihood that the matching segment is IBD (identical by descent) and not IBS (identical by state).”

The same rules for utilizing larger segment data need to be applied to small segment data to begin with.

Are more guidelines needed for small segments?  I don’t know, but we’ll never know if we don’t work with many individual situations and find the common methods for success and identify any problematic areas.

Why Do Small Segments Matter?

In some cases, especially as we work beyond the 6 generation level, small segments may be all we have left of a specific ancestor.  If we don’t learn to recognize and utilize the small segments available to us, those ancestors, genetically speaking, will be lost to us forever.

As we move back in time, the DNA from more distant ancestors will be divided into smaller and smaller segments, so if we ever want the ability to identify and track those segments back in time to a specific ancestor, we have to learn how to utilize small segment data – and if we have deleted that data, then we can’t use it.

In my case, I have identified all of my 5th generation ancestors except one, and I have a strong lead on her.  In my 6th generation, however, I have lots of walls that need to be broken through – and DNA may be the only way I’ll ever do that.

Let’s take a look at what I can expect when trying to match people who also descend from an ancestor 5 generations back in time.  If they are my same generation, they would be my fourth cousins.

Based on the autosomal statistics chart at ISOGG, 4th cousins, on the average, would expect to share about 13.28 cM of DNA from their common ancestor.  This would not be over the match threshold at FTDNA of approximately 20 cM total, and if those segments were broken into three pieces, for example, that cousin would not show as a match at either FTDNA or 23andMe, based on the vendors’ respective thresholds.

% Shared DNA Expected Shared cM Relationship
0.781% 53.13 Third cousins, common ancestor is 4 generations back in time
0.391% 26.56 Third cousins once removed
20 cm Family Tree DNA total cM Threshold
0.195% 13.28 Fourth cousins, common ancestor is 5 generations back in time
7 cM 23andMe individual segment cM match threshold
0.0977% 6.64 Fourth cousins once removed
0.0488% 3.32 Fifth cousins, common ancestor is 6 generations back in time
0.0244 1.66 Fifth cousins once removed

If you’re lucky, as I was with Hickerson, you’ll match at least some relative who carries that ancestral DNA line above the threshold, and then they’ll match other cousins above the threshold, and you can build a comparison network, linking people together, in that fashion.  And yes you may well have to utilize GedMatch for people testing at various different vendors and for those smaller segment comparisons.

For clarification, I have never “called” a genealogy match without supporting large segment data.  At the vendors, you can’t even see matches if they don’t have larger segments – so there is no way to even know you would match below the threshold.

I do think that we may be able to make calls based on small segments, at least in some instances, in the future.  In fact, we have to figure out how to do this or we will rarely be able to move past the 5th or 6th generation utilizing genetics.

At the 5th generation, or third cousins, one expects to see approximately 26 cM of matching DNA, still over the threshold (if divided correctly), but from that point further back in time, the expected shared amount of DNA is under the current day threshold.  For those who wonder why the vendors state that autosomal matches are reliable to about the 5th or 6th generation, this is the answer.

I do not discount small segments without cause.  In other words, I don’t discount small segments unless there is a reason.  Unless they are positively IBS by chance, meaning false, and I can prove it, I don’t disregard them.  I do label them and make appropriate notes.  You can’t learn from what’s not there.

Let me give you an example.  I have one area of my spreadsheet where I have a whole lot of segments, large and small, labeled Acadian.  Why?  Because the Acadians are so intermarried that I can’t begin to sort out the actual ancestor that DNA came from, at least not yet…so today, I just label them “Acadian.”

This example row is from my master spreadsheet.  I have my Mom’s results in my spreadsheet, so I can see easily if someone matches me and Mom both. My rows are pink.  The match is on Mom’s side, which I’ve color coded purple.  I don’t know which ancestor is the most recent common ancestor, but based on the surnames involved, I know they are Acadian.  In some cases, on Acadian matches, I can tell the MRCA and if so, that field is completed as well.

Me Mom acadian

As a note of interest, I inherited my mother’s segment intact, so there was no 50% division in this generation.

I also have segments labeled Mennonite and Brethren.  Perhaps in the future I’ll sort through these matches and actually be able to assign DNA segments to specific ancestors.  Those segments aren’t useless, they just aren’t yet fully analyzed.  As more people test, hopefully, patterns will emerge in many of these DNA groupings, both small and large.

In fact, I talked about DNA patterns and endogamous populations in my recent article, Just One Cousin.

For me, today, some small segment matches appear to be central European matches.  I say “appear to be,” because they are not triangulated.  For me this is rather boring and nondescript – but if this were my African American client who is trying to figure out which line her European ancestry came from, this could be very important.  Maybe she can map these segments to at least a specific ancestral line, which she would find very exciting.

Learning to use small segments effectively has the potential to benefit the following groups of people:

  • People with colonial ancestry, because all that may be left today of colonial ancestors is small segments.
  • People looking to break down brick walls, not just confirm currently known ancestors.
  • People looking for minority ancestors more than 5 or 6 generations back in their trees.
  • Adoptees – although very clearly, they want to work with the largest matches first.
  • People working with ethnic identification of ancestors, because you will eventually be able to track ethnicity identifying segments back in time to the originating ancestor(s).

Conversely, people from highly endogamous groups may not be helped much, if at all, by small segments because they are so likely to be widely shared within that population as a group from a common ancestor much further back in time.  In fact, the definition of a “small segment” for people with fully endogamous families might be much larger than for someone with no known endogamy.

However, if we can identify segments to specific populations, that may help the future accuracy of ethnicity testing.

Let’s go back and take a look at the Hickerson data using the same format we have been using for the comparisons so far.

Small Segment Examples

These Hickerson/Vannoy examples do not utilize random small segment matches, but are utilizing the same matching rules used for larger matches in conjunction with known, triangulated cousin groups from a known ancestor.  Many cousins, including 2 brothers and their uncle all carry this same DNA.  Like in Israel’s case, where did they get that same DNA if not from a common ancestor?

In the following examples, I want to stress that all of the people involved DO HAVE LARGER SEGMENT MATCHES on other chromosomes, which is how we knew they matched in the first place, so we aren’t trying to prove they are a match.  We know they are.  Our goal is to determine if small segments are useful in the same situation, proving matches, as with larger segments.  In other words, do the rules hold true?  And how do we work with the data?  Could we utilize these small segment matches if we didn’t have larger matching segments, and if so, how reliable would they be?

There is a difference between a single match and a triangulated group:

  • Matches between two people are suggestive of a common ancestor but could be IBS by chance or population..
  • Multiple matches, such as with the 6 different Hickersons who descend from Charles Hickerson and Mary Lytle, both in the Ancestry DNA Circle and at Family Tree DNA, are extremely suggestive of a specific common ancestor.
  • Only triangulated groups are proof of a common ancestor, unless the people are  closely related known relatives.

In our Hickerson/Vannoy study, all participants match at least to one other (but not to all other) group members at Family Tree DNA which means they match over the FTDNA threshold of approximately 20 cM total and at least one segment over 7.7cM and 500 SNPs or more.

In the example below, from the Hickerson article, the known Vannoy cousins are on the left side and the Hickerson matches to the Vannoy cousins are across the top.  We have several more now, but this gives you an idea of how the matching stacked up initially.  The two green individuals were proven descendants from Charles Hickerson and Mary Lytle.

vannoy hickerson higginson matrix

The goal here is to see how small data segments stack up in a situation where the relationship is distant.  Can small segments be utilized to prove triangulation?  This is slightly different than in the Just One Cousin article, where the relationship between the individuals was close and previously known.  We can contrast the results of that close relationship and small segments with this more distant connection and small segments.

Sarah Hickerson and Daniel Vannoy

The Vannoy project has a group of about a dozen cousins who descend from Elijah Vannoy who have worked together to discover the identify of Elijah’s parents.  Elijah’s father is one of 4 Vannoy men, all sons of the same man, found in Wilkes County, NC. in the late 1700s.  Elijah Vannoy is 5 generations upstream from me.

What kind of evidence do we have?  In the paper genealogy world, I have ruled out one candidate via a Bible record, and probably a second via census and tax records, but we have little information about the third and fourth candidates – in spite of thoroughly perusing all existent records.  So, if we’re ever going to solve the mystery, short of that much-wished-for Vannoy Bible showing up on e-Bay, it’s going to have to be via genetic genealogy.

In addition to the dozen or so Vannoy cousins who have DNA tested, we found 6 individuals who descend from Sarah Hickerson’s parents, Charles Hickerson and Mary Lytle who match various Vannoy cousins.  Additionally, those cousins match another 21 individuals who carry the Hickerson or derivative surnames, but since we have not proven their Hickerson lineage on paper, I have not utilized any of those additional matches in this analysis.  Of those 26 total matches, at Family Tree DNA, one Hickerson individual matches 3 Vannoy cousins, nine Hickerson descendants match 2 Vannoy cousins and sixteen Hickerson descendants match 1 Vannoy cousin.

Our group of Vannoy cousins matching to the 6 Charles Hickerson/Mary Lytle descendants contains over 60 different clusters of matching DNA data across the 22 chromosomes.  Those 6 individuals are included in 43 different triangulated groups, proving the entire triangulation group shares a common ancestor.  And that is BEFORE we add any GedMatch information.

If that sounds like a lot, it’s not.  Another recent article found 31 clusters among siblings and their first cousin, so 60 clusters among a dozen known Vannoy cousins and half a dozen potential Hickerson cousins isn’t unusual at all.

To be very clear, Sarah Hickerson and Daniel Vannoy were not “declared” to be the parents of Elijah Vannoy, born in 1784, based on small segment matches alone.  Larger segment matches were involved, which is how we saw the matches in the first place.  Furthermore, the matches triangulated.  However, small segments certainly are involved and are more prevalent, of course, than large segments.  Some cousins are only connected by small segments.  Are they valid, and how do we tell?  Sometimes it’s all we have.

Let me give you the classic example of when small segments are needed.

We have four people.  Person A and B are known Vannoy cousins and person C and D are potential Hickerson cousins.  Potential means, in this case, potential cousins to the Vannoys.  The Hickersons already know they both descend from Charles Hickerson and Mary Lytle.

  • Person A matches person C on chromosome 1 over the matching threshold.
  • Person B matches person D on chromosome 2 over the matching threshold.

Both Vannoy cousins match Hickerson cousins, but not the same cousin and not on the same segments at the vendor.  If these were same segment matches, there would be no question because they would be triangulated, but they aren’t.

So, what do we do?  We don’t have access to see if person C and D match each other, and even if we did, they don’t match on the same segments where they match persons A and B, because if they did we’d see them as a match too when we view A and B.

If person A and B don’t match each other at the vendor, we’re flat out of luck and have to move this entire operation to GedMatch, assuming all 4 people have or are willing to download their data.

a and b nomatch

If person A and B match each other at the vendor, we can see their small segment data as compared to each other and to persons C and D, respectively which then gives us the ability to see if A matches C on the same small segment as B matches D.

a and b match

If we are lucky, they will all show a common match on a small segment – meaning that A will match B on a small segment of chromosome 3, for example, and A will match C on that same segment.  In a perfect world, B will also match D on that same segment, and you will have 4 way triangulation – but I’m happy with the required 3 way match to triangulate.

This is exactly what happened in the article, Be Still My H(e)art.  As you can see, three people match on chromosomes 1 and 8, below – two of whom are proven cousins and the third was the wife surname candidate line.

Younger Hart 1-8

The example I showed of chromosome 2 in the Hickerson article was where all participants of the 5 individuals shown on the chromosome browser were matching to the Vannoy participant.  I thought it was a good visual example.  It was just one example of the 60+ clusters of cousin matches between the dozen Vannoy cousins and 6 Hickerson descendants.

This example was criticized by some because it was a small segment match.  I should probably have utilized chromosome 15 or searched for a better long segment example, but the point in my article was only to show how people that match stack up together on the chromosome browser – nothing more.   Here’s the entire chromosome, for clarity.

hickerson vannoy chr 2

Certainly, I don’t want to mislead anyone, including myself.  Furthermore, I dislike being publicly characterized as “wrong” and worse yet, labeled “irresponsible,” so I decided to delve into the depths of the data and work through several different examples to see if small segment data matching holds in various situations.  Let’s see what we found.

Chromosome 15

I selected chromosome 15 to work with because it is a region where a lot of Vannoy descendants match – and because it is a relatively large segment.  If the Hickersons do match the Vannoys, there’s a fairly good change they might match on at least part of that segment.  In other words, it appears to be my best bet due to sheer size and the number of Elijah Vannoy’s descendants who carry this segment.  In addition to the 6 individuals above who matched on chromosome 15, here are an additional 4.  As you can see, chromosome 15 has a lot of potential.

Chrom 15 Vannoy

The spreadsheet below shows the sections of chromosome 15 where cousins match.  Green individuals in the Match column are descendants of Charles Hickerson and Mary Lytle, the parents of Sarah Hickerson.  The balance are Vannoys who match on chromosome 15.

chr 15 matches ftdna v4

As you can see, there are several segments that are quite large, shown in yellow, but there are also many that are under the threshold of 7cM, which are all  segments that would be deleted if you are deleting small segments.  Please also note that if you were deleting small segments, all of the Hickerson matches would be gone from chromosome 15.

Those of you with an eagle eye will already notice that we have two separate segments that have triangulated between the Vannoy cousins and the Hickerson descendants, noted in the left column by yellow and beige.  So really, we could stop right here, because we’ve proven the relationship, but there’s a lot more to learn, so let’s go on.

You Can’t Use What You Can’t See

I need to point something out at this point that is extremely important.

The only reason we see any segment data below the match threshold is because once you match someone on a larger segment at Family Tree DNA, over the threshold, you also get to view the small segment data down to 1cM for your match with that person. 

What this means is that if one person or two people match a Hickerson descendant, for example you will see the small segment data for their individual matches, but not for anyone that doesn’t match the participant over the matching threshold.

What that means in the spreadsheet above, is that the only Hickerson that matches more than one Vannoy (on this segment) is Barbara – so we can see her segment data (down to 1cM ) as compared to Polly and Buster, but not to anyone else.

If we could see the smaller segment data of the other participants as compared to the Hickerson participants, even though they don’t match on a larger segment over the matching threshold, there could potentially be a lot of small segment data that would match – and therefore triangulate on this segment.

This is the perfect example of why I’ve suggested to Family Tree DNA that within projects or in individuals situations, that we be allowed to reduce the match threshold – especially when a specific family line match is suspected.

This is also one of the reasons why people turn to GedMatch, and we’ll do that as well.

What this means, relative to the spreadsheet is that it is, unfortunately, woefully incomplete – and it’s not apples to apples because in some cases we have data under the match threshold, and in some, we don’t.  So, matches DO count, but nonmatches where small segment data is not available do NOT count as a non-match, or as disproof.  It’s only negative proof IF you have the data AND it doesn’t match.

The Vannoys match and triangulate on many segments, so those are irrelevant to this discussion other than when they match to Hickerson DNA.  William (H), descends from two sons of Charles Hickerson and Mary Lytle.  Unfortunately, he only matches one Vannoy, so we can only see his small segments for that one Vannoy individual, William (V).  We don’t know what we are missing as compared to the rest of the Vannoy cousins.

To see William (H)’s and William (V)’s DNA as compared to the rest of the Vannoy cousins, we had to move to GedMatch.

Matching Options

Since we are working with segments that are proven to be Vannoy, and we are trying to prove/disprove if Daniel Vannoy and Sarah Hickerson are the parents of Elijah through multiple Hickerson matches, there are only a few matching options, which are:

  1. The Hickerson individuals will not triangulate with any of the Vannoy DNA, on chromosome 15 or on other chromosomes, meaning that Sarah Hickerson is probably not the mother of Elijah Vannoy, or the common ancestor is too far back in time to discern that match at vendor thresholds.
  2. The Hickerson individuals will not triangulate on this segment, but do triangulate on other segments, meaning that this segment came entirely from the Vannoy side of the family and not the Hickerson side of the family. Therefore, if chromosome 15 does not triangulate, we need to look at other chromosomes.
  3. The Hickerson individuals triangulate with the Vannoy individuals, confirming that Sarah Hickerson is the mother of Elijah Vannoy, or that there is a different common unknown ancestor someplace upstream of several Hickersons and Vannoys.

All of the Vannoy cousins descend from Elijah Vannoy and Lois McNiel, except one, William (V), who descends from the proven son of Sarah Hickerson and Daniel Vannoy, so he would be expected to match at least some Hickerson descendants.  The 6 Hickerson cousins descend from Charles Hickerson and Mary Lytle, Sarah’s parents.

hickerson vannoy pedigree

William (H), the Hickerson cousin who descends from David, brother to Sarah Hickerson, is descended through two of David Hickerson’s sons.

I decided to utilize the same segment “mapping comparison” technique with a spreadsheet that I utilized in the phasing article, because it’s easy to see and visualize.

I have created a matching spreadsheet and labeled the locations on the spreadsheet from 25-100 based on the beginning of the start location of the cluster of matches and the end location of the cluster.

Each individual being compared on the spreadsheet below has a column across the top.  On the chart below, all Hickerson individuals are to the right and are shown with their cells highlighted yellow in the top row.

Below, the entire colorized chart of chromosome 15 is shown, beginning with location 25 and ending with 100, in the left hand column, the area of the Vannoy overlap.  Remember, you can double click on the graphics to enlarge.  The columns in this spreadsheet are not fully expanded below, but they are in the individual examples.

entire chr 15 match ss v4

I am going to step through this spreadsheet, and point out several aspects.

First, I selected Buster, the individual in the group to begin the comparison, because he was one of the closest to the common ancestor, Elijah Vannoy, genealogically, at 4 generations.  So he is the person at Family Tree DNA that everyone is initially compared against.

Everyone who matches Buster has their matching segments shown in blue.  Buster is shown furthest left.

When participants match someone other than Buster, who they match on that segment is typed into their column.  You can tell who Buster matches because their columns are blue on matching locations.  Here’s an example.

Me Buster match

You can see that in my column, it’s blue on all segments which means I match Buster on this entire region.  In addition, there are names of Carl, Dean, William Gedmatch and Billie Gedmatch typed into the cell in the first row which means at that location, in addition to Buster, I also match Carl and Dean at Family Tree DNA and William (descended from the son of Daniel Vannoy and Sarah Hickerson) at Gedmatch and Billie (a Hickerson) at Gedmatch.  Their name is typed into my column, and mine into theirs.  Please note that I did not run everyone against everyone at GedMatch.  I only needed enough data to prove the point and running many comparisons is a long, arduous process even when GedMatch isn’t experiencing problems.

On cells that aren’t colorized blue, the person doesn’t match Buster, but may still match other Vannoy cousin segments.  For example, Dean, below, matches Buster on location 25-29, along with some other cousins.  However, he does not match Buster on location 30 where he instead matches Harold and Carl who also don’t match Buster at that location. Harold, Carl and Dean do, however, all descend from the same son of Elijah so they may well be sharing DNA from a Vannoy wife at this location, especially since no one who doesn’t share that specific wife’s line matches those three at this location.

Me Buster Dean match

Remember, we are not working with random small data segments, but with a proven matching segment to a common Vannoy ancestor, with a group of descendants from a possible/probable Hickerson ancestor that we are trying to prove/disprove.  In other words, you would expect either a lot of Hickerson matches on the same segments, if Hickerson is indeed a Vannoy ancestral family, or virtually none of them to match, if not.

The next thing I’d like to point out is that these are small segments of people who also have larger matching segments, many of whom do triangulate on larger segments on other chromosomes.  What we are trying to discern is whether small segment matches can be utilized by employing the same matching criteria as large segment matching.  In other words, is small segment data valid and useful if it meets the criteria for an IBD match?

For example, let’s look at Daniel.  Daniel’s segments on chromosome 15, were it not for the fact that he matches on larger segments on other chromosomes, would not be shown as matches, because they are not individually over the match threshold.

Look at Daniel’s column for Polly and Warren.

Daniel matches 2

The segments in red show a triangulated group where Daniel and Warren, or Daniel, Warren and Polly match.  The segments where all 3 match are triangulated.

This proves, unquestionably, that small segments DO match utilizing the normal prescribed IBD matching criteria.  This spreadsheet, just for chromosome 15, is full of these examples.

Is there any reason to think that these triangulated matches are not identical by descent?  If they are not IBD, how do all of these people match the same DNA? Chance alone?  How would that be possible?  Two people, yes, maybe, but 3 or more?  In some cases, 5 or 6 on the same segment?  That is simply not possible, or we have disproven the entire foundation that autosomal DNA matching is based upon.

The question will soon be asked if small segments that triangulate can be useful when there are no larger matching segments to put the match over the initial vendor threshold.

Triangulated Groups

As you can see, most of the people and segments on the spreadsheet, certainly the Elijah descendants, are heavily triangulated, meaning that three or more people match each other on the same locations.  Most of this matching is over the vendor threshold at Family Tree DNA.

You can see that Buster, Me, Dean, Carl and Harold all match each other on the same segments, on the left half of the spreadsheet where our names are in each other’s columns.

triangulated groups

Remember when I said that the spreadsheet was incomplete?  This is an example.  David and Warren don’t match each other at a high enough total of segments to get them over the matching threshold when compared to each other, so we can’t see their small segment data as compared to each other.  David matches Buster, but Warren doesn’t, so I can’t even see them both in relationship to a common match.  There are several people who fall into this category.

Let’s select one individual to use as an example.

I’ve chosen the Vannoy cousin, William(V), because his kit has been uploaded to Gedmatch, he has Vannoy matches and because William is proven to descend from Sarah Hickerson and Daniel Vannoy through their son Joel – so we expect some Hickerson DNA to match William(V).

If William (V) matches the Hickersons on the same DNA locations as he matches to Elijah’s descendants, then that proves that Elijah’s descendant’s DNA in that location is Hickerson DNA.

At GedMatch, I compared William(V) with me and then with Dean using a “one to one” comparison at a low threshold, simply because I wanted as much data as I could get.  Family Tree DNA allows for 1 cM and I did the same, allowing 100 SNPs at GedMatch.  Family Tree DNA’s lowest SNP threshold is 500.

In case you were wondering, even though I did lower the GedMatch threshold below the FTDNA minimum, there were 45 segments that were above 1cM and above 500 SNPs when matching me to William(V), which would have been above the lowest match threshold at FTDNA (assuming we were over the initial match threshold.)  In other words, had we not been below the original match threshold (20cM total, one segment over 7.7cM), these segments would have been included at FTDNA as small segments.  As you can see in the chart below, many triangulated.

I colorized the GedMatch matches, where there were no FTDNA matches, in dark red text.  This illustrates graphically just how much is missed when the small segments are ignored in cases with known or probable cousins.  In the green area, the entry that says “Me GedMatch” could not be colorized red (because you can’t colorize only part of the text of a cell) so I added the Gedmatch designation to differentiate between a match through FTDNA and one from GedMatch.  I did the same with all Gedmatch matches, whether colorized or not.

Let’s take a look and see how small segments from GedMatch affect our Hickerson matching.  Note that in the green area, William (V) matches William (H), the Hickerson descendant, and William (V) matches to me and Dean as well.  This triangulates William (V)’s Hickerson DNA and proves that Elijah’s descendants DNA includes proven Hickerson segments.

William (V) gedmatch matches v2

In this next example, I matched William (H), the Hickerson cousin (with no Vannoy heritage) against both Buster and me.

William (H) gedmatch me buster

Without Gedmatch data, only two segments of chromosome 15 are triangulated between Vannoy and Hickerson cousins, because we can’t see the small data segments of the rest of the cousins who don’t match over the threshold.

You can see here that nearly the entire chromosome is triangulated using small segments.  In the chart below, you can see both William(V) and William (H) as they match various Vannoy cousins.  Both triangulate with me.

William V and William H

I did the same thing with the Hickerson descendant, Billie, as compared to both me and Dean, with the same type of results.

The next question would be if chromosome 15 is a pileup area where I have a lot of IBS matches that are really population based matches.  It does not appear to be.  I have identified an area of my chromosomes that may be a pileup area, but chromosome 15 does not carry any of those characteristics.

So by utilizing the small segments at GedMatch for chromosome 15 that we can’t otherwise see, we can triangulate at least some of the Hickerson matches.  I can’t complete this chart, because several individuals have not uploaded to GedMatch.

Why would the Hickerson descendant match so many of the Vannoy segments on chromosome 15?  Because this is not a random sample.  This is a proven Vannoy segment and we are trying to see which parts of this segment are from a potential Hickerson mother or the Vannoy father.  If from the Hickerson mother, then this level of matching is not unexpected.  In fact, it would be expected.  Since we cheated and saw that chromosome 15 was already triangulated at Family Tree DNA, we already knew what to expect.

In the spreadsheet below, I’ve added the 2 GedMatch comparisons, William (V) to me and Dean, and William (H) to me and Buster.  You can see the segments that triangulate, on the left.  We could also build “triangulated groups,” like GedMatch does.  I started to do this, but then stopped because I realized most cells would be colored and you’d have a hard time seeing the individual triangulated segments.  I shifted to triangulating only the individuals who triangulate directly with the Hickerson descendant, William(H), shown in green.  GedMatch data is shown in red.

chr 15 with gedmatch

I would like to make three points.

1.  This still is not a complete spreadsheet where everyone is compared to everyone.  This was selectively compared for two known Hickerson cousins, William (V) who descends from both Vannoys and Hickersos and William (H) who descends only from Hickersons.

2. There are 25 individually triangulated segments to the Hickerson descendant on just this chromosome to the various Vannoy cousins.  That’s proof times 25 to just one Hickerson cousin.

3.  I would NEVER suggest that you select one set of small segments and base a decision on that alone.  This entire exercise has assembled cumulative evidence.  By the same token, if the rules for segment matching hold up under the worst circumstances, where we have an unknown but suspected relationship and the small segments appear to continue to follow the triangulation rules, they could be expected to remain true in much more favorable circumstances.

Might any of these people have random DNA matches that are truly IBS by chance on chromosome 15?  Of course, but the matching rules, just like for larger segments, eliminates them.  According to triangulation rules, if they are IBS by chance, they won’t triangulate.  If they do triangulate, that would confirm that they received the same DNA from a common ancestor.

If this is not true, and they did not receive their common DNA from a common ancestor, then it disproves the fundamental matching rule upon which all autosomal DNA genetic genealogy is based and we all need to throw in the towel and just go and do something else.

Is there some grey area someplace?  I would presume so,  but at this point, I don’t know how to discern or define it, if there is.  I’ve done three in-depth studies on three different families over the past 6 weeks or so, and I’ve yet to find an area (except for endogamous populations that have matches by population) where the guidelines are problematic.  Other researchers may certainly make different discoveries as they do the same kind of studies.  There is always more to be discovered, so we need to keep an open mind.

In this situation, it helps a lot that the Hickerson/Vannoy descendants match and triangulate on larger segments on other chromosomes.  This study was specifically to see if smaller segments would triangulate and obey the rules. We were fortunate to have such a large, apparently “sticky” segment of Vannoy DNA on chromosome 15 to work with.

Does small segment matching matter in most cases, especially when you have larger segments to utilize?  Probably not. Use the largest segments first.  But in some cases, like where you are trying to prove an ancestor who was born in the 1700s, you may desperately need that small segment data in order to triangulate between three people.

Why is this important – critically important?  Because if small segments obey all of the triangulation rules when larger segments are available to “prove” the match, then there is no reason that they couldn’t be utilized, using the same rules of IBD/IBS, when larger segments are not available.  We saw this in Just One Cousin as well.

However, in terms of proof of concept, I don’t know what better proof could possibly be offered, within the standard genetic genealogy proofs where IBD/IBS guidelines are utilized as described in the Phasing article.  Additional examples of small segment proof by triangulation are offered in Just One Cousin, Lazarus – Putting Humpty Dumpty Together Again, and in Demystifying Autosomal DNA Matching.

Raising Elijah Vannoy and Sarah Hickerson from the Dead

As I thought more about this situation, I realized that I was doing an awful lot of spreadsheet heavy lifting when a tool might already be available.  In fact, Israel’s mention of Lazarus made me wonder if there was a way to apply this tool to the situation at hand.

I decided to take a look at the Lazarus tool and here is what the intro said:

Generate ‘pseudo-DNA kits’ based on segments in common with your matches. These ‘pseudo-DNA kits’ can then be used as a surrogate for a common ancestor in other tests on this site. Segments are included for every combination where a match occurs between a kit in group1 and group2.

It’s obvious from further instructions that this is really meant for a parent or grandparent, but the technique should work just the same for more distant relatives.

I decided to try it first just with the descendants of Elijah Vannoy.  At first, I thought that recreated Elijah would include the following DNA:

  • DNA segments from Elijah Vannoy
  • DNA segments from Elijah Vannoy’s wife, Lois McNiel
  • DNA segments that match from Elijah’s descendants spouse’s lines when individuals come from the same descendant line. This means that if three people descend from Joel Vannoy and Phoebe Crumley, Elijah’s son and his wife, that they would match on some DNA from Phoebe, and that there was no way to subtract Phoebe’s DNA.

After working with the Lazarus tool, I realized this is not the case because Lazarus is designed to utilize a group of direct descendants and then compare the DNA of that group to a second group of know relatives, but not descendants.

In other words, if you have a grandson of a man, and his brother.  The DNA shared by the brother and the grandson HAS to be the DNA contributed to that grandson by his grandfather, from their common ancestor, the great grandfather.  So, in our situation above, Phoebe’s DNA is excluded.

The chart below shows the inheritance path for Lazarus matching.

Lazarus inheritance

Because Lazarus is comparing the DNA of Son Doe with Brother Doe – that eliminates any DNA from the brother’s wives, Sarah Spoon or Mary – because those lines are not shared between Brother Doe and Son Doe.  The only shared ancestors that can contribute DNA to both are Father Doe and Methusaleh Fisher.

The Lazarus instructions allow you to enter the direct descendants of the person/couple that you are reconstructing, then a second set of instructions asks for remaining relatives not directly descended, like siblings, parents, cousins, etc. In other words, those that should share DNA through the common ancestor of the person you are recreating.

To recreate Elijah, I entered all of the Vannoy cousins and then entered William (V) as a sibling since he is the proven son of Daniel Vannoy and Sarah Hickerson.

Here is what Lazarus produced.

lazarus elijah 1

Lazarus includes segments of 4cM and 500 SNPs.

The first thing I thought was, “Holy Moly, what happened to chromosome 15?”  I went back and looked, and sure enough, while almost all of the Elijah descendants do match on chromosome 15, William (V), kit 156020, does not match above the Lazarus threshold I selected.  So chromosome 15 is not included.  Finding additional people who are known to be from this Vannoy line and adding them to the “nondescendant” group would probably result in a more complete Elijah.

lazarus elijah 2

Next, to recreate Sarah Hickerson, I added all of the Vannoy cousins plus William (V) as descendants of Sarah Hickerson and then I added just the one Hickerson descendant, William, as a sibling.  William’s ancestor is proven to be the sibling of Sarah.

I didn’t know quite what to expect.

Clearly if the DNA from the Hickerson descendant didn’t match or triangulate with DNA from any of the Vannoy cousins at this higher level, then Sarah Hickerson wasn’t likely Elijah’s mother.  I wanted to see matching, but more, I wanted to see triangulation.

lazarus elijah 3

I was stunned.  Every kit except two had matches, some of significant size.

lazarus elijah 4

lazarus elijah 5 v2

Please note that locations on chromosomes 3, 4 and 13, above, are triangulated in addition to matching between two individuals, which constitutes proof of a common ancestor.  Please also note that if you were throwing away segments below 7cM, you would lose all of the triangulated matches and all but two matches altogether.

Clearly, comparing the Vannoy DNA with the Hickerson DNA produced a significant number of matches including three triangulated segments.

lazarus elijah 6

Where Are We?

I never have, and I never would recommend attempting to utilize random small match segments out of context.  By out of context, I mean simply looking at all of your 1cM segments and suggesting that they are all relevant to your genealogy.  Nope, never have.  Never would.

There is no question that many small segments are IBS by chance or identical by population.  Furthermore, working with small segments in endogamous populations may not be fruitful.

Those are the caveats.  Small segments in the right circumstances are useful.  And we’ve seen several examples of the right circumstances.

Over the past few weeks, we have identified guidelines and tools to work with small segments, and they are the same tools and guidelines we utilize to work with larger segments as well.  The difference is size.  When working with large segments, the fact that they are large serves an a filter for us and we don’t question their authenticity.  With all small segments, we must do the matching and analysis work to prove validity.  Probably not worthwhile if you have larger segments for the same group of people.

Working with the Vannoy data on chromosome 15 is not random, nor is the family from an endogamous population.  That segment was proven to be Vannoy prior to attempts to confirm or disprove the Hickerson connection.  And we’ve gone beyond just matching, we’ve proven the ancestral link by triangulation, including small segments.  We’ve now proven the Hickerson connection about 7 ways to Sunday.  Ok, maybe 7 is an exaggeration, but here is the evidence summed up for the Vannoy/Hickerson study from multiple vendors and tools:

  • Ancestry DNA Circle indicating that multiple Hickerson descendants match me and some that don’t match me, match each other. Not proof, but certainly suggestive of a common ancestor.
  • A total of 26 Hickerson or derivative family name matches to Vannoy cousins at Family Tree DNA. Not proof, but again, very suggestive.
  • 6 Charles Hickerson/Mary Lytle descendants match to Vannoy cousins at Family Tree DNA. Extremely suggestive, needs triangulation.
  • Triangulation of segments between Vannoy and Hickerson cousins at Family Tree DNA. Proof, but in this study we were only looking to determine whether small segment matches constituted proof.
  • Triangulation of multiple Hickerson/Vannoy cousins on chromosome 15 at GedMatch utilizing small segments and one to one matching. More proof.
  • Lazarus, at higher thresholds than the triangulation matching, when creating Sarah Hickerson, still matched 19 segments and triangulated three for a total of 73.2cM when comparing the Hickerson descendant against the Vannoy cousins. Further proof.

So, can small segment matching data be useful? Is there any reason NOT to accept this evidence as valid?

With proper usage, small segment data certainly looks to provide value by judiciously applying exactly the same rules that apply to all DNA matching.  The difference of course being that you don’t really have to think about utilizing those tools with large segment matches.  It’s pretty well a given that a 20cM match is valid, but you can never assume anything about those small segment matches without supporting evidence. So are larger segments easier to use?  Absolutely.

Does that automatically make small segments invalid?  Absolutely not.

In some cases, especially when attempting to break down brick walls more than 5 or 6 generations in the past, small segment data may be all we have available.  We must use it effectively.  How small is too small?  I don’t know.  It appears that size is really not a factor if you strictly adhere to the IBD/IBS guidelines, but at some point, I would think the segments would be so small that just about everyone would match everyone because we are all humans – so the ultimate identical by population scenario.

Segments that don’t match an individual and either or both parents, assuming you have both parents to test, can safely be disregarded unless they are large and then a look at the raw data is in order to see if there is a problem in that area.  These are IBS by chance.  IBS segments by chance also won’t triangulate further up the tree.  They can’t, because they don’t match your parents so they cannot come from an ancestor.  If they don’t come from an ancestor, they can’t possibly match two other people whose DNA comes from that ancestor on that segment.

If both parents aren’t available, or your small segments do match with your parents, I would suggest that you retain your small segments and map them.

You can’t recognize patterns if the data isn’t present and you won’t be able to find that proverbial needle in the haystack that we are all looking for.

Based on what we’ve seen in multiple case studies, I would conclude that small segment data is certainly valid and can play a valid role in a situation where there is a known or suspected relationship.

I would agree that attempting to utilize small segment data outside the context of a larger data match is not optimal, at least not today, although I wish the vendors would provide a way for us to selectively lower our thresholds.  A larger segment match can point the way to smaller segment matches between multiple people that can be triangulated.  In some situations, like the person A, B, C, D Hickerson-Vannoy situation I described earlier in this article, I would like to be able to drop the match threshold to reveal the small segment data when other matches are suggestive of a family relationship.

In the Hickerson situation, having the ability to drop the matching thresholds would have been the key to positively confirming this relationship within the vendor’s data base and not having to utilize third party tools like GedMatch – which require the cooperation of all parties involved to download their raw data files.  Not everyone transferred their data to Gedmatch in my Vannoy group, but enough did that we were able to do what we needed to do.  That isn’t always the case.  In fact, I have an nearly identical situation in another line but my two matches at Ancestry have declined to download their data to Gedmatch.

This not the first time that small segment data has played a successful role in finding genealogy solutions, or confirming what we thought we knew – although in all cases to date, larger segments matched as well – and those larger segment matches were key and what pointed me to the potential match that ultimately involved the usage of the small segments for triangulation.

Using larger data segments as pointers probably won’t be the case forever, especially if we can gain confidence that we can reliably utilize small segments, at least in certain situations.  Specifically, a small segment match may be nothing, but a small segment triangulated match in the context of a genealogical situation seems to abide by all of the genetic genealogy DNA rules.

In fact, a situation just arose in the past couple weeks that does not include larger segments matching at a vendor.

Let’s close this article by discussing this recent scenario.

The Adoptee

An adoptee approached me with matching data from GedMatch which included matches to me, Dean, Carl and Harold on chromosome 15, on segments that overlap, as follows.

adoptee chr 15

On the spreadsheet above, sent to me by the adoptee, we can see some matches but not all matches. I ran the balance of these 4 people at GedMatch and below is the matching chart for the segment of chromosome 15 where the adoptee matches the 4 Vannoy cousins plus William(H), the Hickerson cousin.

  Me Carl Dean Harold Adoptee
Me NA FTDNA FTDNA GedMatch GedMatch
Carl FTDNA NA FTDNA FTDNA GedMatch
Dean FTDNA FTDNA NA FTDNA GedMatch
Harold GedMatch FTDNA FTDNA NA GedMatch
Adoptee GedMatch GedMatch GedMatch GedMatch NA
William (H) GedMatch GedMatch GedMatch GedMatch GedMatch

I decided to take the easy route and just utilize Lazarus again, so I added all of the known Vannoy and Hickerson cousins I utilized in earlier Lazarus calculations at Gedmatch as siblings to our adoptee.  This means that each kit will be compared to the adoptees DNA and matching segments will be reported.  At a threshold of 300 SNPs and 4cM, our adoptee matches at 140cM of common DNA between the various cousins.

adoptee vannoy match

Please note that in addition to matching several of the cousins, our adoptee also triangulates on chromosomes 1, 11, 15, 18, 19 and 21.  The triangulation on chromosome 21 is to two proven Hickerson descendants, so he matches on this line as well.

I reduced the threshold to 4cM and 200 SNPs to see what kind of difference that would make.

adoptee vannoy match low threshold

Our adoptee picked up another triangulation on chromosome 1 and added additional cousins in the chromosome 15 “sticky Vannoy” cluster and the chromosome 18 cluster.

Given what we just showed about chromosome 15, and the discussions about IBD and IBS guidelines and small matching segments, what conclusions would you draw and what would you do?

  1. Tell the adoptee this is invalid because there are no qualifying large match segments that match at the vendors.
  2. Tell the adoptee to throw all of those small segments away, or at least all of the ones below 7cM because they are only small matching segments and utilizing small matching segments is only a folly and the adoptee is only seeing what he wants to see – even though the Vannoy cousins with whom he triangulates are proven, triangulated cousins.
  3. Check to see if the adoptee also matches the other cousins involved, although he does clearly already exceeds the triangulation criteria to declare a common ancestor of 3 proven cousins on a matching segment. This is actually what I did utilizing Lazarus and you just saw the outcome.

If this is a valid match, based on who he does and doesn’t match in terms of the rest of the family, you could very well narrow his line substantially – perhaps by utilizing the various Vannoy wives’ DNA, to an ancestral couple.  Given that our adoptee matches both the Vannoys and the Hickersons, I suspect he is somehow descended from Daniel Vannoy and Sarah Hickerson.

In Conclusion

What is the acceptable level to utilize small segments in a known or suspected match situation?

Rather than look for a magic threshold number, we are much better served to look at reliable methods to determine the difference between DNA passed from our ancestors to us, IBD, and matches by chance.  This helps us to establish the reliability of DNA segments in individual situations we are likely to encounter in our genealogy.  In other words, rather that throw the entire pile of wheat away because there is some percentage of chaff in the wheat, let’s figure out how to sort the wheat from the chaff.

Fortunately, both parental phasing and triangulation eliminate the identical by chance segments.

Clearly, the smaller the segments, even in a known match situation, the more likely they are identical by population, given that they triangulate.  In fact, this is exactly how the Neanderthal and Denisovan genomes have been reconstructed.

Furthermore, given that the Anzick DNA sample is over 12,000 years old, Identical by population must be how Anzick is matching to contemporary humans, because at least some of these people do clearly share a common ancestor with Anzick at some point, long ago – more than 12,000 years ago.  In my case, at least some of the Anzick segments triangulate with my mother’s DNA, so they are not IBS by chance.  That only leaves identical by population or identical by descent, meaning within a genealogical timeframe, and we know that isn’t possible.

There are yet other situations where small segment matches are not IBS by chance nor identical by population.  For example, I have a very hard time believing that the adoptee situation is nothing but chance.  It’s not a folly.  It’s identical by descent as proven by triangulation with 10 different cousins – all on segments below the vendor matching thresholds.

In fact, it’s impossible to match the Vannoy cousins, who are already triangulated individually, by chance.  While the adoptee match is not over the vendor threshold, the segments are not terribly small and they do all triangulate with multiple individuals who also triangulate with larger segments, at the vendors and on different chromosomes.

This adoptee triangulated match, even without the Hickerson-Vannoy study disproves the blanket statement that small segments below 5cM cannot be used for genealogy.  All of these segments are 7.1cM or below and most are below 5.

This small segment match between my mother and her first cousins also disproves that segments under 5cM can never be used for genealogy.

Two cousins combined

This small segment passed from my mother to me disproves that statement too – clearly matching with our cousin, Cheryl.  If I did not receive this from my mother, and she from her parent, then how do we match a common cousin???

me mother small seg

More small segment proof, below, between my mother and her second cousin when Lazarus was reconstructing my mother’s father.

2nd cousin lazarus match

And this Vannoy Hickerson 4 cousin triangulated segment also disproves that 5cM and below cannot be used for genealogy.

vannoy hickerson triang

Where did these small segments come from if not a common ancestor, either one or several generations ago?  If you look at the small segment I inherited from my mother and say, “well, of course that’s valid, you got it from your mother” then the same logic has to apply that she inherited it from her parent.  The same logic then applies that the same small segment, when shared by my mother’s cousin, also came from the their common grandparents.  One cannot be true without the others being true.  It’s the same DNA. I got it from my mother.  And it’s only a 1.46cM segment, shown in the examples above.

Here are my observations and conclusions:

  • As proven with hundreds of examples in this and other articles cited, small segments can be and are inherited from our ancestors and can be utilized for genetic genealogy.
  • There is no line in the sand at 7cM or 5cM at which a segment is viable and useful at 5.1cM and not at 4.9cM.
  • All small segment matches need to be evaluated utilizing the guidelines set forth for IBD versus IBS by chance versus identical by population set forth in the articles titled How Phasing Works and Determining IBD Versus IBS Matches and Demystifying Autosomal DNA Matching.
  • When given a choice, large segment matches are always easier to use because they are seldom IBS by chance and most often IBD.
  • Small segment matches are more likely to be IBS by chance than larger matches, which is why we need to judiciously apply the IBD/IBS Guidelines when attempting to utilize small segment matches.
  • All DNA matches, not just small segments, must be triangulated to prove a common ancestor, unless they are known close relatives, like siblings, first cousins, etc.
  • When working in genetic genealogy, always glean the information from larger matches and assemble that information.  However, when the time comes that you need those small segments because you are working 5, 6 or 7 generations back in time, remember that tools and guidelines exist to use small segments reliably.
  • Do not attempt to use small segments out of context.  This means that if you were to look only at your 1cM matches to unknown people, and you have the ability to triangulate against your parents, most would prove to be IBS by chance.  This is the basis of the argument for why some people delete their small segments.  However, by utilizing parental phasing, phasing against known family members (like uncles, aunts and first cousins) and triangulation, you can identify and salvage the useable small segments – and these segments may be the only remnants of your ancestors more than 5 or 6 generations back that you’ll ever have to work with.  You do not have to throw all of them away simply because some or many small segments, out of context, are IBS by chance.  It doesn’t hurt anything to leave them just sit in your spreadsheet untouched until the day that you need them.

Ultimately, the decision is yours whether you will use small segments or not – and either decision is fine.  However, don’t make the decision based on the belief that small segments under some magic number, like 5cM or 7cM are universally useless.  They aren’t.

Whether small segments are too much work and effort in your individual situation depends on your personal goals for genetic genealogy and on factors like whether or not you descend from an endogamous population.  People’s individual goals and circumstances vary widely.  Some people test at Ancestry and are happy with inferential matching circles and nothing more.  Some people want to wring every tidbit possible out of genealogy, genetic or otherwise.

I hope everyone will begin to look at how they can use small segment data reliably instead of simply discarding all the small segments on the premise that all small segment data is useless because some small segments are not useful.  All unstudied and discarded data is indeed useless, so discarding becomes a self-fulfilling prophecy.

But by far, the worst outcome of throwing perfectly good data away is that you’ll never know what genetic secrets it held for you about your ancestors.  Maybe the DNA of your own Sarah Hickerson is lurking there, just waiting for the right circumstances to be found.

Ancestor Reconstruction

No, this is not Jurassic Park and we’re not actually recreating or cloning our ancestors – just on paper.

Back in early 2012, I began to discuss the possibility of using chromosome mapping of descendants to virtually recreate ancestors.

In 2013, I wrote a white paper about how to do this, and circulated it among a group of scientists who I was hoping would take the ball and run, creating tools for genetic genealogists.  So far, that hasn’t happened, but what has happened is that I’ve adapted a tool created by Kitty Cooper for something entirely different than its original purpose to do a “proof of concept.”

Kitty Cooper created the Ancestor Chromosome Mapper to allow people to map the DNA contributed by different ancestors on their chromosomes.  It’s exciting to see your ancestors mapped out, in color, on your chromosomes.

I utilized Kitty’s tool, found here, to map the proven DNA of my ancestors, below, utilizing autosomal matching and triangulation, to create this ancestor map of my own chromosomes.  As you can see there are still a lot of blank spaces.

Roberta's ancestor map2

After thinking about this a bit, I realized that I could do the same thing for my ancestors.

The chromosomes shown would be those of an individual ancestor, and the DNA mapped onto the chromosomes would be from the proven descendants that they inherited from that ancestor.  Eventually, with enough descendants we could create a “virtual file” for that ancestor to represent themselves in autosomal matching.  So, one day, I might create, or find created by someone else, a DNA “recreated” file for Abraham Estes, born in 1647 in Nonington, Kent, or for Henry Bolton, born about 1760 in England, or any of my other ancestors – all from the DNA of their descendants.

I decided a while back to take this concept for a test spin.

I wanted to see a visual of Joseph Preston Bolton’s DNA on his chromosomes, and who carries it today.  I wrote about this in Joseph’s 52 Ancestors article.

Utilizing Kitty Cooper’s wonderful ancestor chromosome mapping tool, a little differently than she had in mind, I mapped Joseph’s DNA and the contributors are listed to the right of his chromosome.  You can build a virtual ancestor from their descendants based on common matching segments, so long as they don’t share other ancestral lines as well.  I have only utilized the proven, or triangulated DNA segments, where three people match on the same segment.

joseph bolton reconstructed

We have a couple more DNA testers that descended from Joseph Bolton’s father, Henry Bolton through children other than Joseph Preston Bolton.  Adding these segments to the chromosome chart generated for Joseph Preston Bolton, we see the confirmed Henry Bolton segments below.

henry bolton proven

On the chart above, I’ve only used proven segments.

On the next chart I have not been able to “prove” all of the segments through triangulation (3 people), but if all of the provisional segments are indeed Bolton segments, then Henry’s chromosome map would have a few more colored segments.  Clearly, we need a lot more people to test to create more color on Henry’s map, but still, it’s pretty amazing that we can recreate this much of Henry’s chromosome map from these few descendants.

henry bolton probably

There’s a lot of promise in this technique.  Henry Bolton was married twice.  By looking at the DNA the two groups of children, 21 in total, have in common, we know that their common DNA comes from Henry himself.  DNA that is shared between only the groups descended from first wife, Catherine Chapman, but not from second wife, Nancy Mann, or vice versa, would be attributed to the wife of the couple.  Since Henry was married twice, with enough testers, it would be possible to reconstruct, in part, at least some of the genome of both wives, in addition to Henry.

Now, think for a minute, a bit further out in time.

We don’t know who Nancy Mann’s parents are for sure, although we’ve done a lot of eliminating and we know, probably, who her father was, and likely who her grandfather and great-grandfather were….but certainty is not within grasp right now.

But, it will be in the future through ancestor reconstruction.

Let’s say that the descendants of John Mann, the immigrant, reconstruct his genome.  He had 4 known sons and they had several children, so that would be possible.  John, the immigrant, is believed to be Nancy’s great-grandfather through son John Jr.

Now, let’s say that some of those segments that we can attribute through Henry Bolton’s children, as described above, are attributable to Nancy Mann.  The X chromosome match above is positively Nancy’s DNA.  How do I know that?  because it came through her son, Joseph Preston Bolton, and men don’t inherit an X chromosome from their father, only their mother.  So today, 3 descendants carry that segment of Nancy Mann’s X chromosome.

Let’s say that one of the Nancy Mann’s proven DNA segments (not the X, because John didn’t give his X to his son John) matches smack dab in the middle of one of the proven “John Mann” segments.  We’ve just proven that indeed, Nancy is related to John.

Think about the power of this for adoptees, for those who don’t know who their parent or parents are for other reasons, and for those of us who have dead end brick walls who are wives with no surnames.  Who doesn’t have those?

We have the potential, within the foreseeable future, to create “ancestor libraries” that we can match to in order to identify our ancestors.  Once the ancestor is reconstructed, kind of like reconstituting something dehydrated with water, we’ll be able to utilize their autosomal DNA file to make very interesting discoveries about them and their lives.  For example, eye color – at GedMatch today there is an eye color predictor.  There are several ethnicity admixture tools.  Want to know if your ancestor was ethnically admixed?  Virtually recreate them and find out.

Once recreated, we will be able to discover hair color, skin color and all of the other traits and medical conditions that we can today discover through the trait testing at Family Tree DNA and the genetic predispositions that Promethease reveals.

Yes, there will be challenges, like who creates those libraries, moderates any disputes and where are they archived for comparison….but those are details that can be worked out.  Maybe that’s one of the new roles of project administrators or maybe we’ll have ancestor administrators.

Someday, it may be possible to construct an entire family tree from your DNA combined with proven genealogy trees – not by intensely laborious work like it’s done today, but with the click of a button.

And that someday is very likely within our lifetimes, and hopefully, shortly.  The technology and techniques are here to do it today.

I surely hope one of the vendors implements this functionality, and soon, because, like all genealogists, I have a list of genealogy mysteries that need to be solved!!!

Big Y Release

Drum roll…the big day is finally here.

Family Tree DNA held a webinar meeting today to explain the new Big Y product features for a number of us who blog or otherwise educate within the genetic genealogy community.

First, the results will begin rolling today, not tomorrow.  100 will initially be released today and the balance of the initial orders will be released as they finish QA over the next month, at which point, Family Tree DNA anticipates their backlog will be resolved.  There were thousands of tests ordered.  They aren’t saying how many thousands.

First, a little background.  There are 36,562 known Y SNPs in the Family Tree DNA data base that everyone is being compared to.  In the example we saw of the delivered product, 25,749 has been found and callable at a high confidence rate in the individual being tested and were reported.  Low confidence calls are not reported on this personal delivery page, but are included in the data download files.

Big Y landing

On the customer’s personal page, there are two tabs.  The first Tab is for reporting against known SNPs.

Y page 1 cropped

The second is for Novel Variations, in other words, SNPs not on the list of 36,562 known and previously named SNPs.

Y page 2

In essence, Family Tree DNA has implemented a 4 step process.

  1. An individual’s sequenced data is compared to the SNP data base and divided into two categories, known and previously unknown.  The customer’s data is delivered based on these two categories.
  2. All customer data is being loaded into a mammoth size data base at which point it will be determined which SNPs (please see the definition of a SNP here) are actually undiscovered SNPs that will be named, and which are truly novel, family or clan variants.
  3. New SNPs that are found in enough of the population will be named and will be added to the haplotree.
  4. Novel variants will remain that, and will continue to be reported on client pages.

Family Tree DNA is still working on items 2-4.  In addition, they are working on a white paper which will be out in the next 6 weeks or so that will discuss things like the average number of novel SNPs per person being discovered, mutation rates, performance metrics and cross validation of platforms between the next gen sequencing Illumina equipment, Sanger sequencing and chip based sequencing, like the Geno 2.0 chip.

What’s Being Reported?

According to Dr. David Mittelman, the Y chromosome has about 60 million letters.  About half of those are inverted repeats and are therefore not sequenceable.

Of the balance, there are several with poor readability, for example, some that simulate the X, etc.  These are also not useful or reliable to read.

That leaves about 10 million, these being the gold standard of Y sequencing.  Family Tree DNA tries to read about 13.5 million of these base pairs.  They promised 10 million positions when they announced this product.  They are delivering between 11.5 and 12.5 million positions per person.  They also promised about 25,000 common variants, meaning known SNPs and they are delivering between 25,000 and 30,000 per person.  This is only counting medium to high confidence calls.  The low confidence calls are included in the download files, but not counted in this total or shown on your personal page.

Exactly how many locations are reported for any individual are shown on the bottom left hand side of the page.  This example is generic.  Yours might say something like, “Showing 1 of 10 of 25,000 of 36,564.”  In this case, 25,000 would be the number of SNPs read and called on your test.

Big Y total

All 25,000 or so results are being shown, both positive and negative.  That way, there is no question about whether a specific location was tested, or the outcome.  Of course, the third and fourth outcome options are a no-call or poor confidence call at that location.

All novel mutations are being reported by reference number so that they can be compared to like data from any source, as opposed to an “in-house” assigned number.

Insertions and deletions are also in the download files, but not reported on the customer’s delivery page.

Personal data is also searchable by SNP.

SNP search

Individual SNP Testing

After steps 2 and 3 have occurred, it has to be determined which SNPs are found in a high enough percentage of a population to warrant primer development to test individual SNP positions.

Family Tree DNA also clarified something from the November conference.  The 2000 SNP limit is only how many SNPs can be loaded at one time, not the total number they will ever develop primers for or test for.  They will do what makes sense in terms of the SNP being present in enough of the market to warrant primer development.  With the very large number of Novel SNPs being discovered, it wouldn’t make much sense to purchase 50 individual SNP tests at $39 each.  The break even point today, at $39, would be 17 individual SNPs, as compared to the $695 Big Y test.  I expect that eventually the demand for individual SNP testing will decrease substantially.

Downloadable Files

Available on everyone’s page is the ability to download 2 files, a VCF (variant call file) which lists the variants identified as compared to the human reference sequence and the BED file which is a text file which shows a range of positions that passed the QC.

They will also be making available the BAM raw data files within the next week or so, but are finalizing the delivery methodology due to the very large file sizes involved.

The Much Anticipated HaploTree

If I had a dollar for every time someone has asked when the new tree would be available, I’d be a rich woman.  As we all know, there have been a couple of problems with the tree.  The new tree is 7 to 8 times the size of the 2010 tree.  The tree, of course, has been cast in warm jello, an ever-moving target.  And with the SNP tsumani that has been arriving with the full sequencing of the Y chromosome, that tree will very shortly be much larger still.

Bennett Greenspan said today that an updated tree is, “Needed, desired and will be delivered.”  He went on to say that they have had two teams working together with Nat Geo for the past couple of months to both finalize the tree itself and to work on the customer interface.  Since the tree is much larger, it’s not as easy as the older trees which could be seen at a glance and easily navigated.  Furthermore, there is also the matter of integration with National Geographic.

Bennett says an updated tree will be delivered “within the next several weeks.”

New SNPs that are discerned to be SNPs and not novel/clan or family variations will then be named and added to the tree.

Integration

The initial release of Big Y data will be just that, a release of the results of the data, displayable on your personal page and downloadable.  The newly found SNPs will not initially update the current haplotree on your personal page.  This is the same issue we have today with the transfer and integration of Nat Geo data, because the tree is not current, so this is nothing new.  The implementation of the new tree however, will remedy both problems.

The Future

Never happy with what we have, genetic genealogists will want a way to match to other people on SNPs, just like we do today with STR markers.  In fact, we’ll want a way to integrate that matching and discern what it means to our own private family or clan situations.

Family Tree DNA is aware of that, planning for it, and welcomes feedback for how they can make this information even more useful in the future than it is today.

New Orders

I expect this delivery of new information via Big Y results will indeed spur a new interest in ordering this test from people who were waiting to see exactly what was being delivered.  For those people ordering now, they can expect an 8-10 week turnaround, so long as additional vials aren’t required for testing.

For More Information

Elise Friedman is holding the free Big Y Webinar tomorrow, Friday, February 28th.  You can read about it, sign up and learn how to access this and other webinars after their initial showing at this link.

Family Tree DNA FAQ pages you’ll want to visit are here and here.

Generational Inheritance

Autosomal DNA testing has opened up the brave new world for genealogists.  Along with that opportunity comes some amount of frustration and sometimes desperation to wring every possible tidbit of information out of autosomal results, sometimes resulting in pushing the envelope of what the technology and DNA can tell us.

I often have clients who want me to take a look at DNA results from people several generations removed from each other and try to determine if the ancestors are likely to be brothers, for example.  While that’s fairly feasible in the first few generations, the further back in time one goes, the less reliably we can say much of anything about how DNA is transmitted.  Hence, the less we can say, reliably, about relationships between people.

The best we can ever do is to talk in averages.  It’s like a coin flip.  Take a coin out right now and flip it 10 times.  I just did, and did not get 5 heads and 5 tails, which the average would predict.  But averages are comprised of a large number of outcomes divided by the actual number of events.  That isn’t the same thing as saying if one repeats the event 10 times that you will have 5 heads and 5 tails, or the average.  Each of those 10 flips are entirely independent, so you could have any of 11 different outcomes:

  • 0 heads 10 tails
  • 1 head 9 tails
  • 2 heads 8 tails
  • 3 heads 7 tails
  • 4 heads 6 tails
  • 5 heads 5 tails
  • 6 heads 4 tails
  • 7 heads 3 tails
  • 8 heads 2 tails
  • 9 heads 1 tail
  • 10 heads 0 tails

What the average does say is that in the end, you are most likely to have an average of 5 heads and 5 tails – and the larger the series of events, the more likely you are to reach that average.

My 10 single event flips were 4 heads and 6 tails, clearly not the average.  But if I did 10 series of coin flips, I bet my average would be 5 and 5 – and at 100 flips, it’s almost assured to be 50-50 – because the population, or number of events, has increased to the point where the average is almost assured.

You can see above, that while the average does indeed map to 5-5, or the 50-50 rule, the results of the individual flips are no respecter of that rule and are not connected to the final average outcome.  For example, if one set of flips is entirely tails and one set of flips is entirely heads, the average is still 50/50 which is not at all reflective of the actual events.

And so it goes with inheritance too.

However, we have come to expect that the 50% rule applies most of the time.  We knowriffle shuffle that it does, absolutely, with parents.  We do receive 50% of our DNA from each parent, but which 50%?.  From there, it can vary, meaning that we don’t necessarily get 25% of each grandparent’s DNA.  So while we receive 50% in total from each parent, we don’t necessarily receive every other segment or location, so it’s not like a rifle card shuffle where every other card is interspersed.

If one parents DNA sequence is:

TACGTACGTACG

A child cannot be presumed to receive every other allele, shown in red below.

TACGTACGTACG

The child could receive any portion of this particular segment, all of it, or none of it.

So, if you don’t receive every other allele from a parent, then how do you receive your DNA and how does that 50% division happen?  The bottom line is that we don’t know, but we are learning.  This article is the result of a learning experience.

Over time, genetic genealogists have come to expect that we are most likely to receive 25% of our DNA from each grandparent – which is statistically true when there are enough inheritance events.  This reflects our expectation of the standard deviation, where about 2/3rds of the results will be within the closest 25% in either direction of the center.  You can see expected standard deviation here.

This means that I would expect an inheritance frequency chart to look like this.

expected inheritance frequency

In this graph above, about half of the time, we inherit 50% of the DNA of any particular segment, and the rest of the time we inherit some different amount, with the most frequently inherited amounts being closer to the 50% mark and the outliers being increasingly rare as you approach 0% and 100% of a particular segment.

But does this predictability hold when we’re not talking about hundreds of events….when we’re not talking about population genetics….but our own family genetics, meaning one transmission event, from parent to child?  Because if that expected 50% factor doesn’t hold true, then that affects DRAMATICALLY what we can say about how related we are to someone 5 or 6 generations ago and how can we analyze individual chromosome data.

I have been uncomfortable with this situation for some time now, and the increasing incidence of anecdotal evidence has caused me to become increasingly more uncomfortable.

There are repeated anecdotal instances of significant segments that “hold” intact for many generations.  Statistically, this should not happen.  When this does happen, we, as genetic genealogists, consider ourselves lucky to be one of the 1% at the end of spectrum, that genetic karma has smiled upon us.  But is that true?  Are we at the lucky 1% end of the spectrum?

This phenomenon is shown clearly in the Vannoy project where 5 cousins who descend from Elijah Vannoy born in 1786 share a very significant portion of chromosome 15.  These people are all 5 generations or more distantly related from the common ancestor, (approximate 4th cousins) and should share less than 1% of their DNA in total, and certainly no large, unbroken segments.   As you can see, below, that’s not the case.  We don’t know why or how some DNA clumps together like this and is transmitted in complete (or nearly complete) segments, but they obviously are.  We often call these “sticky segments” for lack of a better term.

cousin 1

I downloaded this chromosome 15 information into a spreadsheet where I can sort it by chromosome.  Below you can see the segments on chromosome 15 where these cousins match me.

cousin 2

Chromosome 15 is a total of 141 cM in length and has 17,269 SNPs.  Therefore, at 5 generations removed, we would expect to see these people share a total of 4.4cM and 540 SNPs, or less for those more distantly related.  This would be under the matching threshold at either Family Tree DNA or 23andMe, so they would not be shown as matches at all.  Clearly, this isn’t the case for these 5 cousins.  This DNA held together and was passed intact for a total of 25 different individual inheritance events (5 cousins times 5 events, or  generations, each.)  I wrote about this in the article titles “Why Are My Predicted Cousin Relationships Wrong?”

Finally, I had a client who just would not accept no for an answer, wanted desperately to know the genetically projected relationship between two men who lived in the 1700s, and I felt an obligation to look into generational inheritance further.

About this same time, I had been working with my own matches at 23andMe.  Two of my children have tested there as well, a son and a daughter, so all of my matches at 23andMe obviously match me, and may or may not match my children.  This presented the perfect opportunity to study the amount of DNA transmitted in each inheritance event between me and both children.

Utilizing the reports at www.dnagedcom.com, I was able to download all of my matches into a spreadsheet, but then to also download all of the people on my match list that all of my matches match too.

I know, that was a tongue twister.  Maybe an example will help.

I match John Doe.  My match list looks like this and goes on for 353 lines.

match list

I only match John Doe on one chromosome at one location.  But finding who else on my match list of 353 people that John Doe matches is important because it gives me clues as to who is related to whom and descends from the same ancestor.  This is especially true if you recognize some of the people that your match matches, like your first cousin, for example.  This suggests, below that John Doe is related to me through the same ancestor as my first cousin, especially if John matches me with even more people who share that ancestor.   If my cousin and I both match John Doe on the same segment, that is strongly suggestive that this segment comes from a common ancestor, like in the previous Vannoy example.

Therefore, I methodically went through and downloaded every single one of my matches matches (from my match list) to see who was also on their list, and built myself a large spreadsheet.  That spreadsheet exercise is a topic for another article.  The important thing about this process is that how much DNA each of my children match with John Doe tells me exactly how much of my DNA each of my children inherited from me, versus their father, in that segment of DNA.

match comparison

In the above example, I match John Doe on Chromosome 11 from 37,000,000-63,000,000.  Looking at the expected 50% inheritance, or normal distribution, both of my children should match John Doe at half of that.  But look at what happened.  Both of my children inherited almost exactly all of the same DNA that I had to give.  Both of them inherited just slightly less in terms of genetic distance (cM) and also in terms of the number of SNPs.

It’s this type of information that has made me increasingly skeptical about the 50% bell curve standard deviation rule as applied to individual, not population, genetics.  The bell curve, of course, implies that the 50% percentile is the most likely even to occur, with the 49th being next most likely, etc.

This does not seem to be holding true.  In fact, in this one example alone, we have two examples of nearly 100% of the data being passed, not 50% in each inheritance event.  This is the type of one-off anecdotal evidence that has been making me increasingly uncomfortable.

I wanted something more than anecdotal evidence.  I copied all of the match information for myself and my children with my matches to one spreadsheet.  There are two genetic measures that can be utilized, centimorgans (cM) or total SNPs. I am using cM for these examples unless I state otherwise.

In total, there were 594 inheritance events shown as matches between me and others, and those same others and my children.

Upon further analysis of those inheritance events, 6 of them were actually not inheritance events from me.  In other words, those people matched me and my children on different chromosomes.  This means that the matches to my children were not through me, but from their father’s side or were IBS, Inherited by State.

son daughter comparison

This first chart is extremely interesting.  Including all inheritance events, 55% of the time, my children received none of the DNA I had to give them.  Whoa Nellie.  That is not what I expected to see.  They “should have” received half of my DNA, but instead, half of the time, they received none.

The balance of the time, they received some of my DNA 23% of the time and all of my DNA 21% of the time.  That also is not what I expected to see.

Furthermore, there is only one inheritance event in which one of my children actually inherited exactly half of what I had to offer, so significantly less than 1% at .1%.  In other words, what we expected to see actually happened the least often and was vanishingly rare when not looking at averages but at actual inheritance events.

Let’s talk about that “none” figure for a minute.  In this case, none isn’t really accurate, but I can’t be more accurate.  None means that 23andMe showed no match.  Their threshold for matching is 7cM (genetic distance) and 700 SNPS for the first matching segment, and then 5cM and 700 SNPS for secondary matching segments.  However, if you have over 1000 matches, which I do, matches begin to “fall off,” the smallest ones first, so you can’t tell what the functional match threshold is for you or for the people you match.  We can only guess, based on their published thresholds.

So let’s look at this another way.

Of the 329 times that my children received none of my DNA, 105 of those transmissions would be expected to be under the 700cM threshold, based on a 50% calculation of how many cMs I matched with the individual.  However, not all of those expected events were actually under the threshold, and many transmissions that were not expected to be under that threshold, were.  Therefore, 224, or 68% of those “none” events were not expected if you look at how much of my DNA the child would be expected to inherit at 50%.

Another very interesting anomaly that pops right up is the number of cases where my children inherited more than I had to give them.  In the example below, you can see that I match Jane Doe with 15.2cM and 2859 SNPs, but my daughter matches Jane with 16.3cM and 2960 on the same chromosome.

spreadsheet layout

There are a few possibilities to explain this:

  • My daughter also matches this person on her father’s side at this transition point.
  • My daughter matches this person IBS at this point.
  • The 23andMe matching software is trying to compensate for misreads.
  • There are misreads or no calls in my file.

There of course may be a combination of several of these factors, but the most likely is the fact that she is IBS at this location and the matching software is trying to be generous to compensate for possible no-calls and misreads.  I suggest this because they are almost uniformly very small amounts.

Therefore when my children match me at 100% or greater, I simply counted it as an exact match.  I was surprised at how many of these instances there were.  Most were just slightly over the value of 2 in the “times expected” column.  To explain how this column functions, a value of 1 is the expected amount – or 50% of my DNA.  A value of 2 means that the child inherited all of the DNA I had to offer in that location.  Any value over 2 means that one or more of the bulleted possibilities above occurred.

Between both of my children, there were a total of 75, or 60% with values greater than 2 on cMs and 96, or 80%, on SNPs, meaning that my children matched those people on more DNA at that location than I had to offer.  The range was from 2 to 2.4 with the exception of one match that was at 3.7.  That one could well be a valid transition (other parent) match.

There has been a lot of discussion recently about X chromosome inheritance.  In this case, the X would be like any other chromosome, since I have two Xs to recombine and give to my children, so I did not remove X matches from these calculations.  The X is shown as chromosome 99 here and 23 on the graphs to enable correct column sorting/graphing.

In the chart below, inheritance events are charted by chromosome.  The “Total” columns are the combined events of both my son and daughter.  The blue and pink columns are the inheritance events for both of them, which equal the total, of course.

The “none” column reflects transmissions on that chromosome where my children received none of my DNA.  The “some” column reflects transmission events where my children received some portion of my DNA between 0 (none) and 100% (all).  The “all” column reflects events where my children received all of the DNA that I had to offer.

chromosomal comparison

I graphed these events.

total inheritance graph

The graph shows the total inheritance events between both of my children by chromosome.  Number 23 in these charts is the X chromosome.

son inheritance graph

daughter inheritance graph

These inheritance numbers cause me to wonder what is going on with chromosome 5 in the case of both my daughter and son, and also chromosome 6 with my son.  I wonder if this would be uniform across families relative to chromosome 5, or if it is simply an anomaly within my family inheritance events.  It seems odd that the same anomaly would occur with both children.

son daughter inheritance graph

What this shows is that we are not dealing with a distribution curve where the majority of the events are at the 50% level and those that are not are progressively nearer to the 50% level than either end.  In other words, the Expected Inheritance Frequency is not what was found.

expected inheritance frequency

The actual curve, based on the inheritance events observed here, is shown below, where every event that was over the value of 2, or 100%, was normalized to 2.  This graph is dramatically different than the expected frequency, above.

actual inheritance frequency

Looking at this, it becomes immediately evident that we inherit either all of nothing of our parents DNA segments 85% of the time, and only about 15% of the time we inherit only a portion of our parents DNA segments.  Very, very rarely is the portion we inherit actually 50%, one tenth of one percent of the time.

Now that we understand that individual generational inheritance is not a 50-50 bell curve event, what does this mean to us as genetic genealogists?

I asked fellow genetic genealogist, Dr. David Pike, a mathematician to look this over and he offered the following commentary:

“As relationships get more distant, the number of blocks of DNA that are likely to be shared diminishes greatly.  Once down to one block, then really there are three outcomes for subsequent inheritance:  either the block is passed intact, no part of it is passed on, or recombination happens and a portion of it is passed on.  If we ignore this recombination effect (which should rarely affect a small block) then the block is either passed on in an “all or nothing” manner.  There’s essentially no middle ground with small blocks and even with lots of examples it doesn’t really make sense to expect an average of 50%.  As an analogy, consider the human population:  with about half of us being female and about half of us being male, the “average” person should therefore be androgynous, and yet very few people are indeed androgynous.”

In other words, even if you do have a segment that is 10 cMs in length, it’s not 10 coin flips, it’s one coin flip and it’s going to either be all, nothing or a portion thereof, and it’s more than 6 times more likely to be all or nothing than to be a partial inheritance.

So how do we resolve the fact that when we are looking at the 700,000 or so locations tested at Family Tree DNA and the 600,000 locations tested at 23andMe, that we can in fact use the averages to predict relationships, at least in closely related individuals, but we can’t utilize that same methodology in these types of individual situations?  There are many inheritance events being taken into consideration, 600,000 – 700,000, an amount that is mathematically high enough to over overcome the individual inheritance issues.  In other words, at this level, we can utilize averages.  However, when we move past the larger population model, the individual model simply doesn’t fit anymore for individual event inheritance – in other words, looking at individual segments.

Dr. Pike was kind enough to explain this in mathematical terms, but ones that the rest of us can understand:

“I think that part of what is at stake is the distinction between continuous versus discrete events.  These are mathematical terms, so to illustrate with an example, the number line from 0 to 10 is continuous and includes *all* numbers between them, such 2.55, pi, etc.  A discrete model, however, would involve only a finite number of elements, such as just the eleven integers from 0 to 10 inclusive.  In the discrete model there is nothing “in between” consecutive elements (such as 3 and 4), whereas in the continuous model there are infinitely elements between them.

It’s not unlike comparing a whole spectrum against a finite handful of a few options.  In some cases the distinction is easily blurred, such as if you conduct a survey and ask people to rate a politician on a discrete scale of 0 to 10… in this case it makes intuitive sense to say that the politician’s average rating was 7.32 (for example) even though 7.32 was not one of the options within the discrete scale.

In the realm of DNA, suppose that cousins Alice and Bob share 9 blocks of DNA with each other and we ask how many blocks Alice is likely to share with Bob’s unborn son.  The answer is discrete, and with each block having a roughly 50/50 chance we expect that there will likely be 4 or 5 blocks shared by Alice and Bob Jr., although the randomness of it could result in anywhere from 0 to 9 of the blocks being shared.  Although it doesn’t make practical sense to say that “four and a half” blocks will likely be shared [well, unless we allow recombination to split a block and thereby produce a shared “half block”], there is still some intuitive comfort in saying that 4.5 is the average of what we would expect, but in reality, either 4 or 5 blocks are shared.

But when we get to the extreme situation of there being only 1 block, for which the discrete options are only 0 or 1 block shared, yes or no, our comfortable familiarity with the continuous model fails us.  There are lots of analogies here, such as what is the average of a coin toss, what is the average answer to a True/False question, what the average gender of the population, etc.

Discrete models with lots of options can serve as good approximations of continuous situations, and vice-versa, which is probably part of what’s to blame for confusion here.

Really DNA inheritance is discrete, but with very many possible segments [such as if we divided the genome up into 10 cM segments and asked how many of Alice’s paternal segments will be inherited by one of her children, we can get away with a continuous model and essentially say that the answer is roughly 50%.  Really though, if there are 3000 of these blocks, the actual answer is one of the integers:  0, 1, 2, …, 2999, 3000.  The reality is discrete even though we like the continuous model for predicting it.

However, discrete situations with very few options simply cannot be modelled continuously.”

Back to our situation where we are attempting to determine a relationship of 2 men born in the 1700s whose descendants share fragments of DNA today.  When we see a particularly large fragment of DNA, we can’t make any assumptions about age or how long it has been in existence by “reverse engineering” it’s path to a common ancestor by doubling the amount of DNA in every generation.  In other words, based on the evidence we see above, it has most likely been passed entirely intact, not divided.  In the case of the Vannoy DNA, it looks like the ends have been shaved a few times, but the majority of the segment was passed entirely intact.  In fact, you can’t double the DNA inherited by each individual 5 times, because in at least one case, Buster, doubling his total matching cM, 100, even once would yield a number of cM greater the size of chromosome 15 at 141 cM.

Conversely, when we see no DNA matches, for example, in people who “should be” distant cousins, we can’t draw any conclusions about that either.  If the DNA didn’t get passed in the first generation – and according to the numbers we just saw – 58% doesn’t get passed at all, and 26% gets passed in its entirety, leaving only about 15% to receive some portion of one parent’s DNA, which is uniformly NOT 50% except for one instance in almost 1000 events (.1%) – then all bets for subsequent generations are off – they can’t inherit their half if their half is already gone or wasn’t half to begin with.

Based on mathematical model, Probability of Recombination, Dr. Pike has this to say:

If I’m reading this right, a 10 cM block has a 10% chance of being split into parts during the recombination process of a single conception. Although 10% is not completely negligible, it’s small enough that we can essentially consider “all” or “nothing” as the two dominant outcomes.

This is the fundamental underlying reason why testing companies are hesitant to predict specific relationships – they typically predict ranges of relationships – 1st to 3rd cousin, for example, based on a combination of averages – of the percentages of DNA shared, the number of segments, the size of segments, the number of SNPs etc.  The testing company, of course, can have no knowledge of how our individual DNA is or was actually passed, meaning how much ancestral DNA we do or don’t receive, so they must rely on those averages, which are very reliable as a continuous population model, and apparently, much less so as discrete individual events.

I would suggest that while we certainly have a large enough sample of inheritance events between me and my two children to be statistically relevant, it’s not large enough study to draw any broad sweeping conclusions. It is, after all, only 3 people and we don’t know how this data might hold up compared to a much larger sample of family inheritance events.  I’d like to see 100 or 1000 of these types of studies.

I would be very interested to see how this information holds up for anyone else who would be willing to do the same type of information download of their data for parent/multiple sibling inheritance.  I will gladly make my spreadsheet with the calculations available as a template to anyone who wants to do the same type of study.

I wonder if we would see certain chromosomes that always have higher or lower generational inheritance factors, like the “none” spike we see on chromosome 5.  I wonder if we would see a consistent pattern of male or female children inheriting more or less (all or none) from their parents.  I wonder what other kinds of information would reveal itself in a larger study, and if it would enable us to “weight” match information by chromosome or chromosome/gender, further refining our ability to understand our genetic relationships and to more accurately predict relationships.

I want to thank Dr. David Pike for reviewing and assisting with this article and in particular, for being infinitely patient and making the application of the math to genetics understandable for non-mathematicians.  If you would like to see an example of Dr. Pike’s professional work, here is one of his papers.  You can find his personal web page here and his wonderful DNA analysis tools here.