I’ve recently had the opportunity to perform two, 4-generation, inheritance studies.
In both of these cases, we have the DNA of 4 generations: grandmother, parent, child and grandchild or grandchildren. I’ll be using the second study because there are two great-grandchildren to compare.
Let me introduce you to the players.
I wanted, with real data, to address some assertions and assumptions that I see being made periodically in the genetic genealogy community. We need to know if these hold up to scrutiny, or not. Besides that, it’s just fun to see what happens to DNA with 4 generations and 5 people to compare.
What kinds of information are we looking to confirm or refute in this study?
1 – That small segments don’t occur within a couple generations, meaning that that DNA can’t be or isn’t broken into small segments that quickly.
2 – That small segments can never be used genealogically and are not useful.
3 – That DNA is most of the time passed in 50% packages. While this is true in the first generation, meaning a child does receive half of each parent’s DNA, they do not receive 25% of each grandparent’s DNA.
4 – That segments over a certain threshold, like 5 or 7 cM, are all reliable as IBD (identical by descent.)
5 – That segments under a certain threshold, like 5 or 7 cM are all unreliable and should never be used, in fact, cannot ever be used and should be discarded.
6 – That there is a rule that you cannot have more than two crossovers per chromosome.
All individuals tested at Family Tree DNA and we’ll be using the FTDNA chromosome browser for comparisons.
First, let’s look at the amount of expected DNA matching versus the actual amount of DNA matching, per generation. The entire number of cM being measured is 6766.2, per the ISOGG Autosomal Statistics Wiki page.
Expected vs Actual Inheritance Chart
This chart compares the expected versus actual amount of DNA shared between person 1 and person 2,
|Person 1||Person 2||Expected DNA Match cM/%||Actual DNA Match|
|Grandmother||Parent (grandmother’s child)||3383.1 / 50%||3384.03 / 50.01%|
|Grandmother||Pink Child (grandmother’s grandchild)||1691.5 / 25%||1670.64 / 24.69%|
|Grandmother||Blue Grandchild (grandmother’s great-grandchild)||845.775 / 12.5%||704.84 / 10.39%|
|Grandmother||Green Grandchild (grandmother’s great-grandchild)||845.775 / 12.5%||842.64 / 12.45%|
Now, let’s take a look at our chromosome data. Keep in mind, everyone is being compared to the oldest generation – in this case – the great-grandmother’s DNA.
- The background chromosome belongs to the great-grandmother of the youngest generation – meaning everyone is being compared to her.
- Grandparent = orange – because the child receives 50% of each parent’s DNA, the orange child of the great-grandmother will match her DNA 100%.
- Grandchild = pink – since the grandchild is being compared to the grandparent, and not their parent, we will see how much of the grandmother’s DNA the pink child received. The dark spaces are the “ghost image” of the grandfather’s DNA – identified by the lack of the grandmother’s DNA in that location.
- Oldest great grandchild = blue
- Youngest great grandchild = green
The two great grandchildren are full siblings. None of the parents involved are related to each other or to other generational spouses. This has been confirmed both by genealogy pedigree chart and by utilizing the tools at GedMatch for comparisons to each other as well as the “are your parents related” tool.
The first comparison, below, shows the 4 individuals compared to the great grandmother’s DNA at the Family Tree DNA with the match default set at 5cM
The image below, shows the same individuals after dropping the match criteria to 1cM. Several small colored segments appear.
I downloaded all of the matching data for these individuals into a spreadsheet so that I could work with the actual chromosomal data. I’m not boring you with that here, but I have used the raw matching data for the actual comparisons.
Let’s talk about what a crossover is, because understanding crossovers are important
Crossover example 1 – A crossover is where you start/stop receiving DNA from one grandparent or the other. This is easy to see if we look at chromosome 1.
In this example, the parent is orange and the child is pink but they are both being compared to the grandparent of the pink person, the mother of the orange person.
What this means is that while the orange person will always match the grey background chromosome of their mother, the pink person will only match their grandmother on the portion of the DNA they received from their mother that was from their grandmother. The pink person received their grandfather’s DNA in some locations, and not their grandmother’s. Where that transition happens is called a crossover and it is where the colored segment stops, as noted by the arrows above, and the back background begins, indicating no match to the grandmother.
You can see that the matches span the center of the chromosome where the grey area indicates there is no data being read. There is also a second small grey area to the right of the center. Ignore these grey areas. They are in essence DNA deserts where there isn’t enough DNA to be read or useful. Family Tree DNA (and other vendors) stitch the data on both sides together, so to speak, and matches on both sides of this area are considered to be contiguous matches.
You can see that the pink person has two crossover areas where they stopped receiving DNA from the mother’s mother (background chromosome being compared against) and instead started receiving DNA from the mother’s father. How do we know that? There only two people who contributed the orange parent’s DNA that the pink child inherited. If the pink child did not inherit the orange parent’s Mom’s DNA on this segment, then the pink child had to have inherited the orange parent’s Dad’s DNA.
Crossover example 2 – A second kind of crossover is where you are still receiving DNA from the same parent, but from different ancestors on that parental line
I’ve created a chart to illustrate this phenomenon
The names in the charts at the bottom are the people who tested today. All of these individuals are known cousins who are from my mother’s side. The name at the top is the common ancestor of all of the testers.
In the first situation, in locations 1-5, Me, Charlie and David match. None of the three of us match our cousin, Mary on those locations. However, moving to locations 6-10, Me, Charlie and Mary match each other, but not David. Looking at our pedigree charts, we can see that the cousins are matching on different ancestral lines.
Me, Charlie and David share a wife’s line, Sally (wife of John), that Mary does not share. Me, Charlie and Mary share common DNA from George, a male further upstream in that line. George’s son John married Sally. Mary descends from George through a different child, which is why she does not match any of us on the segments we received from Sally, John’s wife.
If you’re just looking at the question, “do Charlie and I match?” the answer would of course be yes, but until we look at a broader spectrum of cousins, we won’t know that our match is actually from two different people in the same descendancy line and that we have an ancestor crossover between locations 5 and 6. However, we’re still receiving our DNA from the same parent, but which ancestor of that parent contributed the DNA has switched
How prevalent are crossovers?
Number of Crossover Events
These are all parent/child crossovers where the DNA donor switched. We can only determine that this happened because we can compare generationally against the grey background great grandmother to the youngest generation
- Orange parent to Pink child – 49
- Pink child to Blue child – 47
- Pink child to Green child – 39
The most segmented chromosome, chromosome 1, has 5 separate matching segments for the blue great grandchild (as compared to the great-grandmother), or 10 crossover events (because neither end was at the beginning or end, although start and end numbers are sometimes “fuzzy”). You can see where a crossover event occurs when the DNA goes from matching to non-matching.
I downloaded all of our matching data into a spreadsheet so that I can work with the segment matches individually.
Looking at the data, there are a few things that jump out immediately:
- On chromosomes 4 and 14, the pink child received none of the orange grandmother’s DNA. That means that the pink child had to have received the grandfather’s DNA for all of chromosome 15. So, if anyone thinks that the 50% rule really works uniformly across generations – here’s concrete proof that it doesn’t. Furthermore, this occurred for an entire chromosome – twice out of 23 chromosomes, or 8.7% of the time.
- On chromosome 11, the exact opposite happened. The pink child received all of the grandmother’s chromosome, but barely gave any to their blue child. The blue child received their mother’s DNA in that location. On chromosome 13, the pink child received almost all of the grandmother’s DNA.
- Please note that while the averages of expected versus inherited DNA work out pretty closely, when averaging across all 23 chromosomes, as shown in the Expected vs Actual Inheritance Chart, the individual chromosomes and how much of which grandparent’s or great-grandparent’s DNA is inherited varies wildly from none to 100%.
- There are several locations on 10 different chromosomes where the DNA has been passed generationally intact 2 or 3 times, without division.
- Several small segments have been created within 3 transmission events.There are small green and blue segments on several different chromosomes which reflect very small amounts of the great grandmother’s DNA inherited by the green and blue great-grandchildren. This conclusively dismisses the theory that small segments aren’t ever created within a couple of generations.
- Chromosome 10 is very choppy, including small blue and green grandchild segments that match the orange grandparent and the great-grandmother without having matches to the pink child. This means that those unconnected blue and green small segments are either identical by chance or there is a read issue with the pink person’s DNA on this chromosome.
- There are a total of 31 small segments, meaning under 7cM. Of those, a total of 10 do not triangulate, meaning they match the grandmother but they do not match their parent. The 7 pink segments appear to triangulate, but without another generation of transmission (like the blue and green great-grandchildren), or without the grandfather’s DNA, or without triangulation with a known relative on that segment, it’s impossible to tell for sure. Therefore, 14, or 45% are valid segments and do triangulate.
- There are a total of 92 chromosomal transmission events that took place, meaning that 23 chromosomes got passed from the background person to their orange child, 23 from the orange child to their pink child, 23 from the pink child to the blue grandchild and 23 from the pink child to the green grandchild.
- Furthermore, based on this limited study, at least 32.26% of the small segments do not triangulate and are not IBD, but are instead identical by chance.
- In three instances, the exact DNA (from the great grandmother) was given to both the green and blue great grandchildren. In eight other events, the same DNA, without division, was given from a parent to one child.
- There are several instances, on chromosomes 3, 4, 9, 14, 15, 16, 20, and 22 where the pink child passed none of their grandmother’s DNA to their child, even though they inherited the grandmother’s DNA.
Individual Chromosomes and Their Messages
I’d like to walk through several chromosomes and chat a little bit about what we’re seeing.
First, I’d like to illustrate the difference between chromosome matches at the default level (the first chromosome, above) and at the 1cM level (the lower chromosome.) At the lower match threshold, you will see additional small segment matches that are not shown at the higher threshold, noted by red arrows.
Let’s take a look at the messages held by our individual chromosomes.
On all of these chromosomes, you’ll see that the orange child matches thier mother, the background person being compared against, exactly, on every location that is measured. Half of everyone’s DNA comes from their mother, so all of their DNA will match to her on any given chromosome. Remember, we are only measuring matching DNA (half identical segments) – so the other half of the person’s DNA that matches their father is not shown.
I have left the orange segments in the graphics, even though they all match on the entire chromosome length, so you can see the continuity from generation to generation. Pink is the orange person’s child, so you can see that the pink child inherited part of the DNA the orange person inherited from their mother, but not all. The part that is black in the pink row, as compared to the orange segment, means that the pink child inherited that DNA from their grandfather at those locations – and not the grandmother being compared against
In one instance, on chromosome 1, the pink child gave their grandmother’s DNA to both of their children. You can see that to the far left with the red arrow.
You can also see that the blue grandchild only received a small part of their great grandmother’s DNA, but the green grandchild received a much larger segment.
In one area, the pink child clearly received their grandmother’s DNA, but didn’t give any of it to either the blue or green grandchild, shown below at the red arrow. There is no blue or green matching the great-grandmother’s DNA.
To the right of the arrow, top, above, you can see where the pink child contributed their grandmother’s DNA to their blue child, but not to the green child. The pink child contributed their other parent’s DNA in that instance, bottom, above, because their child does not match their orange mother – so that DNA had to come from the grandfather.
On the chromosome match that includes the smaller segments, below, you can see there are a total of 5 segments not shown with the higher threshold.
The first two arrows, on the left, point to small segments shared by the blue and green grandchildren with their great-grandmother and their pink parent – so these triangulate and they are fine.
The third arrow, on the right hand side pointing to the green segment that does not match with the pink parent indicates a match that is identical by chance. We’ll talk more about this in chromosome 3.
The fourth arrow, at the far right, shows a small segment of orange DNA that was passed to their pink child, but the pink child did not pass it on to either of their children. This segment could be a legitimate segment by descent, but it could also be by chance. We’ll talk about that more on chromosome 8.
Chromosome 2 shows two small segments. You can see that the pink child gave a significant portion of their grandmother’s DNA to the blue child, but only two small segments to the green child in that region, at the red arrows. They do triangulate though, because they match their parents. See how nicely the DNA stacks up between all of the generations.
The pink child inherited very little of the grandmother’s DNA in this region. Of the small amount the pink child did inherit, the pink child gave even less of it to their children. One small piece to the green grandchild, shown at right, and none to the blue grandchild.
Why, then, is there a lonely blue segment on this comparison chromosome showing that the blue great-grandchild matches their orange grandmother and their great-grandmother, but not their pink parent? This is the first example of an identical by chance segment (or a read error in the pink parent’s file).
Three Kinds of DNA Match Segments
There are three kinds of DNA segment matches.
- Identical by descent (IBD) where you receive the segment from your ancestors and we can track it as far back up the tree as we have living people. This is the example where the small segment of the great-grandchildren (blue or green) match their parent (pink), their grandparent (orange) and their great-grandmother’s background chromosome being compared against.
- Identical by state (IBS) which sometimes is used to mean not identical by descent. What it actually means is that you can still match and receive the DNA from your ancestors, but the segment may be very prevalent in a specific community or ethnic group. An alternative explanation is that the DNA ‘state’ is so common that everyone in that area has it, so it’s virtually useless in identifying ancestors, because you can’t really tell which lines it came from. So IBS does triangulate, because it did come from a common ancestor, but you may match a large number of people at this location. Portions of chromosome 6 are known to fall into this category. More often than not, I hear IBS used to indicate that there is a match, but the common ancestor isn’t known or hasn’t yet been identified.
- Identical by chance (IBC) is where a specific DNA combination is a match, but it’s not a match because it was handed down ancestrally, but simply by the luck of the draw. Because everyone carries the DNA of both parents, sometimes people can match you by zigzagging back and forth between your father’s and mother’s DNA. These matches aren’t ancestral, but just by luck or chance. Shorter matches, meaning small segments, are much more likely to be identical by chance than longer matches. When you have both parents DNA, you can easily eliminate IBC segments because they won’t triangulate – as we have just demonstrated on chromosome 3.
Chromosome 4 is particularly interesting because the orange person matches their background mother, of course, but apparently their pink child inherited this entire chromosome from the pink person’s grandfather – because the pink person does not match their grandmother – there are no pink matching segments to the background grandmother.
On chromosome 5, the pink child matches the grandmother on almost the entire chromosome, except for a small part to the left of center.
You may notice that there is a segment of blue that appears to extend beyond the pink bar at the left arrow – which would mean that the blue area matches the great-grandmother without matching the pink parent. The segments on the chromosome map are not exactly to scale, and the beginnings and ends are sometimes what is referred to as fuzzy. This means that they are not exact measurements but that they in essence the absence or presence of DNA in a bucket of a specific size. If any part of your DNA is in that bucket, then your start or stop segment are the edges of that bucket. In this case, the entire match is 47.51cM for the pink child and 49.82 for the blue grandchild, so the difference may or may not be relevant.
Although this actually is a small matching segment, or non-matching segment, you would never notice this if you were just looking at the blue grandchild matching to the great grandmother. It’s only with the introduction of the parent’s pink DNA that you notice that the blue great grandchild’s DNA match with the great grandmother extends beyond that of the parent.
Chromosome 6 is rather unremarkable except that the orange person seems to have had a read or file error of some sort. The orange results are shown in two separate pieces, but we know that the orange person must match their mother 100%. We know this issue is in the orange person’s file, because their pink child and both of the blue and green grandchildren match the background person, the orange persons’ mother, with no break in their DNA.
Chromosome 7 shows another example of 5 generations matching with the stacking of orange, blue, green and pink against the background person’s chromosome, at right. It also shows another example an identical by chance match, with the blue grandchild showing a match to their great-grandmother but no match to their pink parents, near the center at the red arrow.
Chromosome 8 shows another example of the pink child having inherited a small segment of their grandmother’s DNA, but not passing it on to their children.
How do we know if this is a legitimate IBD segment, or if it something else? Since the pink child will match their mother 100%, and they didn’t pass it on tho their children, how can we prove that the small pink segment where they match their grandmother is IBD.
How could we prove this one way or the other?
First of all, it probably doesn’t matter, except as a matter of interest – or unless of course this one segment is THE one you need to identify that colonial ancestor. If this was a normal match, we could just see if the match matched the child and the parent too, which would immediately phase the match against their parent – but we can’t do that when matching to a grandparent because the child will always match their parent 100%.
If you have the grandfather’s DNA at Family Tree DNA, you could compare the pink grandchild to their grandfather. On chromosome 8, the grandfather’s DNA in the pink row is identified by the dark grey – because it’s where the pink grandchild does not match their grandmother – so they must match their grandfather on that segment because their orange parent only had two pieces of DNA to give them, the piece from their mother or the piece from their father.
Therefore, if this is a valid segment, then you won’t see at match in the grandfather’s DNA on same portion of the segment. If you see a match to both the grandmother and the grandfather, it’s likely that the small segment match to the grandmother is not identical by descent – you but really don’t know for sure.
How could that be? I asked David Pike that question and he pointed out that in one case, he discovered that the grandparents both shared the same DNA segment. The child inherited it from one parent or the other, and passed it on to their child, but since the mother’s and father’s DNA was identical, there is no way to tell which grandparent the segment actually came from. And in this case, the segment would match both grandparents. That is a trait of endogamy and of IBS, or identical by population. If you’re saying, BOO, HISS, about now, I totally understand.
After talking to David, I also realized that if your DNA at those locations just happens to be all homozygous, for example, all Ts, on both sides, for a run of SNPs in a row, and if your parents and grandparents have Ts in either location, you will match them…and anyone else who does too.
So here we have an example of a match that could be IBD if it truly is a small segment by descent and you don’t match the other grandparent at that location. It could be IBC or IBS (by population) if you match both of your grandparents on this segment – but it might be IBD. It’s IBD from one and IBC/IBS from the other – but which one is which?
However, since I don’t have the grandfather’s DNA at Family Tree DNA, my only other alternative is to move to GedMatch and create a phased kit for the grandfather by subtracting the grandmother’s DNA from her orange child, which will give me the DNA the orange child received from their father. Then I can compare the pink grandchild to the grandfather’s phased kit – which is the father’s DNA that the orange child received. This is fine, even if it is only half of the grandfather’s DNA – it s the half that the pink child’s mother received and passed a portion to the pink child.
I would suggest doing this entire exercise on either Family Tree DNA or on the GedMatch platform, and not jumping back and forth between the two. The start and stop segments aren’t exactly the same, and sometimes the segments read differently, creating more segments at GedMatch than at FTDNA. I’m not saying that is wrong, just that it isn’t consistent between the two platforms and when you are dealing with small segments, in particular, you need consistency.
On chromosome 9, the pink child received little of the grandmother’s DNA, and gave none of it to their green child. And yes, if you have a good eye the blue child’s right boundary is slightly beyond the their pink parents – so – you already know what that means. Either a fuzzy boundary or a slight piece of DNA that happened to match with the great-grandmother identical by chance (IBC.)
This chromosome is incredibly interesting because it’s comprised of all small segments. In fact, this is the exact reason why you NEED to look at the 1cM range. At the default setting, if there are no matches except the orange person to their mother. It looks like none of the grandmother’s DNA was passed to the pink child, but in fact, may not be the case. There are three segments passed to the pink child, although the pink child did not pass these on to either of their children. See the discussion on segment 8 about how to tell for sure, if you need to.
The blue and green segments, since they do not match their pink parent are not IBD but are instead IBC. The really interesting part of this is that in one case, the blue and green grandchildren’s DNA matches the orange grandmother on the same segments exactly, but does not match the pink parent.
How can this possible be, you ask, barring a file read issue? Good question. Remember, each child inherits half of their parent’s DNA. In this case, both children apparently inherited the same DNA from both parents, but it wasn’t the orange DNA, but that of the pink child’s father.
It just happened, when the blue and green children’s DNA combined with that of their mother, it just happens to read as a match, for a small segment. You can read about how this might happen in the article, “How Phasing Works and Determining IBD Versus IBS Matches.”
Unfortunately, all these comparisons can do is to tell us simply what does and does not match – they can’t tell us why. Sometimes, based on other comparisons, like phasing and triangulation, we can figure out the “why” part of the puzzle – and sometimes, we can’t.
On chromosome 11, the pink child inherited all of the grandmother’s DNA through their orange parent, but gave less than half to their green child and a small segment to the blue child. The pink child gave the exact same segment in the center to both their blue and green children.
On chromosome 12, the pink child inherited little of their grandmother’s DNA, but passed every bit of what they inherited to both of their children, shown by the nice stack at right. The start and stop locations are exact between the three.
However, in addition, we have three small segments where the green and blue grandchildren match their orange grandmother without matching their pink parent – so those are IBC.
The pink child inherited almost all of their grandmother’s entire chromosome, except for a very small bit at the far right end. The pink child passed almost their entire chromosome 13 to their green child, but only a small amount to the blue child.
This story is easy. The pink child inherited their grandfather’s entire chromosome 14 because they do not match their grandmother’s DNA at all.
This is a very “normal” chromosome. The pink child inherited about half of their grandmother’s DNA and gave about half of what they inherited to their green child. Of course, their blue child got left out altogether – but that looks to be a lot more “normal” than we once thought.
I am skipping chromosome 16-22, because they are more of what you’ve already seen and is, by now, quite familiar Plus, you can take a look at the full chromosome comparison graphic and do your own analysis.
The X chromosome is a bit different, and I’d like to take a look at that.
The X chromosome has special inheritance properties that other chromosomes don’t have. In particular, women inherit an X just like they inherit their other chromosomes from 1-22 – one from Mom and one from Dad. Men, however, only receive an X from their mother. Therefore, there are relatives that you cannot inherit any X DNA from. I wrote about this here and here along with examples and charts.
In this example, the inheritance path is such that it does not affect what can and cannot be inherited since we are comparing to a great-grandmother, but in other situations, this would not be the case.
One last observation about the X chromosome. I have found matching on the X to be particularly unreliable, and have found several situations, where, due to those special inheritance properties, we know beyond any doubt that the common ancestor on the X cannot be the same ancestor as has triangulated on the other chromosomes. So word to the wise – be very vigilant and hesitant to draw conclusions from X matching. I never utilize the X without corroborating autosomal matches and even then, I’m very reticent.
On the average, we do inherit about half of our DNA from in each generation from each ancestral generation. But the average and the actuality of what happens is two entirely different things. Averages are made up of all of the outliers, and if you are one of those outliers, the average isn’t really relevant to you. Kind of reminds me of “one size fits all” which really means “one size fits almost nobody well” and “everyone is some shade of unhappy.”
I wrote about generational inheritance and how it doesn’t always work the way we think, or expect. It’s very important to pay close attention to your own DNA and not rely on averages unless you have absolutely no other choice – and only then understanding the averages are likely wrong in one direction or the other – but it’s the best we’ve got, under the circumstances.
So what can we apply to our genealogy from this little experiment.
- Some of the small segments across 4 generations are valid, meaning identical by descent or IBD.
- At least one third of the small segments aren’t valid and are identical by chance, or IBC.
- Without some form of triangulation or parental phasing, it’s impossible to tell which small segments are and are not valid, or identical by descent.
- Small segments are indeed formed within a 2 or 3 generation span, so they are not always a results of many generations of dividing.
- However, the further back in time your ancestor, the more likely that they will only be represented in your DNA by small segments, if any.
- Many small segments are valid and are not a result of IBC. However, most are not and one needs to understand how to recognize signs of an IBC vs an IBD match.
- Disregarding small segments uniformly is like throwing away the only clues you may have to your most distant ancestors – which are likely your brick walls.
- The largest segment that was not valid was 3.14cM and 600 SNPs.
- The smallest valid segment was 1.25cM and 500 SNPs.
Getting the Most Out of Your DNA Experience
There is a lot more information available to us in our DNA results than is first apparent. It takes a bit of digging and you need to understand how autosomal DNA works in order to ferret out those secrets. Don’t discount or ignore evidence because it’s more difficult to use – meaning small segments. The very piece or breadcrumb you need to solve a long-standing mystery may indeed be right there waiting for you. Learn how to use your DNA information effectively and accurately – including those small segments.
You need to test every cousin you can find and convince to swab or spit. It’s those cousin matches that help immensely with triangulation and confirming the validity of all DNA segments, matching them back to common ancestors. You are building walkways or maybe pathways back in time, with your DNA as the steppingstones. Genetic genealogy is not a one person endeavor. It takes a village, hopefully of cousins willing to DNA test!