What Does MCRA (MRCA) Really Mean??

The MCRA or time to the Most Common Recent Ancestor is a calculation provided by both Family Tree DNA and Ancestry.com for their clients who have taken the Y-line DNA tests.  This is also written MRCA, Most Recent Common Ancestor, MCA and all of the above prefaced with a T meaning “time to”.  Regardless of which way you see this acronym, it means the same thing – the closest ancestor you share with someone in the DNA line being tested.

I have a great example of how this actually translates into reality using the results from both companies.

Often, I receive communications from people who say something like this:

“It says that I’m related to John Doe within 6 generations.  I have both of our genealogies to 6 generations, and I can’t find our common ancestor.  What is wrong?”

The answer to “What is wrong?” is easy.  The person doesn’t understand what the tool that estimates MCRA is telling them.  And, I’m betting they didn’t read the instructions and explanations either, that is if they tested at Family Tree DNA who provides such.

Family Tree DNA provides a great deal more information and a far more robust tool than Ancestry.  Family Tree DNA begins with this information:

“The probability that John Doe and William Doe shared a common ancestor within the last X generations…”  The number of generations and the percentage probability are shown below.

You can also change the generational display.  I changed mine to “every generation.”

This is followed by an explanation and instructions for how to refine the calculations:

Refine your results with paper trail input

However, these results can be refined if their paper trail indicates that no common ancestor between John Doe and William Doe could have lived in a certain number of past generations.

If you don’t know this information for a fact, do not change the “1” in the box in the next paragraph. However, if you have the information, please enter in the box and click on the recalculate button.

John Doe and William Doe did not share a common ancestor more recently than 1 generation(s). (Because the important factor in calculating the time to the Most Recent Common Ancestor is the number of generations between which mutations could take place, the number of years per generation is irrelevant in FTDNATiP™ calculations).

After that, additional explanation and a reference to a FAQ sheet:

* The FTDNATiP™ results are based on the mutation rate study presented during the 1st International Conference on Genetic Genealogy, on Oct. 30, 2004. The above probabilities take into consideration the mutation rates for each individual marker being compared.

Since each marker has a different mutation rate, identical Genetic Distances will not necessarily yield the same probabilities. In other words, even though John Doe has a Genetic Distance‡ of 4 from William Doe, someone else with the same Genetic Distance may have different probabilities, because the distance of 4 was prompted by mutations in different markers, with different mutation rates.

‡Note: The Genetic Distance is the count of the total difference between two individuals. For example, if a marker differs by 2, then the Genetic Distance will count this as a distance of 2.

More questions? Please refer to the FTDNATiP™ FAQ page.

This is a huge difference compared to Ancestry who only gives you a number with absolutely no explanation at all:

The MCRA is the small number beside the name – so John Doe is an MCRA of 2 and William is 24.  I have highlighted these in red below so that you can see them.

Here is the explanation the Ancestry which is followed by the match table.

“You could be close to a meaningful family connection! The list below is sorted by how close your DNA matches (MRCA). The closest matches are at the top.”

Real Life Example

Ok, but what does all this really mean, in real life, to me?

Fortunately, I have a client who has tested at both locations, and has another man who he matches both at Ancestry and at Family Tree DNA.  In addition, we know who their common ancestor is, and we can use this information to compare the accuracy and usefulness of the MCRA calculations.

At Ancestry, these men have tested 34 markers in common and have 4 mutations difference.  Ancestry calls this relationship a distant match at 24 generations to the most common recent ancestor (MCRA).

At Family Tree DNA, they have tested 37 markers in common and have 4 mutations.  Family Tree DNA, without refining the MCRA with the paper trail, calls this as the 50th percentile at 11 generations.  This means that there is a 50% chance that you have a common ancestor within 11 generations.  I use the 50th percentile number because that is the “most likely” spot – meaning that it’s equally likely that your ancestor was closer generationally or further away.

We know that these men are at 8 generations to a common ancestor for one man and 7 generations for the other.

Checking Family Tree DNA’s chart for 7 and 8 generations, that percentage or probability is 20% and 27% respectively.

Interestingly enough, Family Tree DNA says that at 24 generations, which was Ancestry’s estimated number of generations, there is a 97+% likihood that indeed they have a common ancestor.

So what we’ve learned is that Ancestry, aside from providing no tools or explanation, is very, very conservative.  In this case, the number they give you is more likely their 100% sure number, not their “most likely” 50th percentile number.  In fact, if we divide their number in half, it’s still high.

We’ve learned that Family Tree DNA’s 50th percentile is much closer to reality, even without any tweaking that you can do based on known pedigree charts.

23 thoughts on “What Does MCRA (MRCA) Really Mean??

  1. Are they more conservative with their reported autosomal matches too? Makes me wonder if my predicted 3rd cousin match on Ancestry is actually closer…..!

    • It will be interesting to see a comparison like this if someone has a match with the same person at both companies. If someone does come across this situation and would be willing to have their results used as an example for a posting, please let me know.

      • I have done autosomal testing with all three companies and I would be happy to share for an example if you tell me what you need. However, because Ancestry folks don’t always use their real names, I have not been able to tell if any of my 23andMe or FTDNA matches are also matching me on Ancestry. I suspect there are a few but I do not know for sure. I did write one Ancestry match last night who I think, by his Ancestry username, may also be a match with me on FTDNA, but he has not replied yet. If they are the same person, FTDNA has him as a predicted 4th cousin match, and Ancestry, 5th-8th.

        Unfortunately I cannot confirm the actual relationships on paper because I am an adoptee.

      • Jim and I matched with Ancestry as 99% 1st cousins. Jim had no family tree at all. Match was made on DNA only. Jim joined FTDNA to match my 1st cousin. Unfortunately, my 1st cousin is unique and does not match with anyone. FTDNA has done Y-DNA67 and Y-HAP-Backbone in 3/5/2014. They still have nothing on my cousin. Back to Jim and I, I would like to upload my Ancestry to FTDNA to see if we are matched the same. On GEDmatch, we match as high with a MRCA of 1.5. Someone want to help me upload from GEDmatch or Ancestry to FTDNA?

  2. Pingback: Y DNA – Family Tree DNA vs Ancestry | DNAeXplained – Genetic Genealogy

  3. Pingback: Surprise Y Matches – What do they Mean? | DNAeXplained – Genetic Genealogy

  4. In your example above, Ancestry.com isn’t really being more conservative, it is reporting the calculated number of generations with a higher confidence level (i.e. 97+ %) rather than reporting the number of generations to the 50% confidence level. I don’t use the 50% confidence level myself. Statisticians typically use the 95% confidence level. Another hint is to eliminate the first 3 or 4 generations as possibilities if you know those family lines well and the individual to whom you match was not a member of your most recent 3 or 4 generations of family lines.

    In FTDNA’s autosomal testing (they branded it their “Family Finder” test), they are too optimistic in their relationship calculations. Almost all of the matches that I’ve verified so far were at least one generation earlier than FTDNA estimated.

    • FTDNA said I was second cousins with a lady. It turns out our MRCA was born in 1628. That was 11 FTDNA generations off. They consider beyond 7 generations “distant” which is how most of my results are listed which could mean anywhere in time.

      • Suzanne: Then you have a more recent common ancestor with the lady than you realize, and just haven’t found it yet. What if her tree is off in one spot with a hidden adoption? Bottom line, though, is that there is no way you could share enough DNA to be called 2nd cousins at FTDNA, and have had all that DNA come from a MRCA that far back. My guess is that none of that shared DNA came from the ancestor you identified.

      • Are you suggesting a new relative coming back into the family tree? Did I say that right?
        In other words a branch in the tree returned back in.

  5. Pingback: Mitochondrial DNA Smartmatching – The Rest of the Story | DNAeXplained – Genetic Genealogy

  6. Pingback: 2013’s Dynamic Dozen – Top Genetic Genealogy Happenings | DNAeXplained – Genetic Genealogy

  7. Do you have any suggestions for proving a very ancient MRCA? My genealogical data takes me back to 1400 fully confirmed with primary data records. DNA confirms a MRCA back to 1581. I believe the MRCA for my 1400 ancestor will be found in a well documented distant cousin branch where that MRCA would be 36 FTDNA generations ago. There were three potential parents and when I found a DNA project for one linking to that ancient MRCA I was happy to find results. FTDNA predictions only go to 24 generations and so any discussions back that far seem to be confusing. I have 10/12, 11/12, 12/12 up to 29/37 y-str matches to three surnames known to come from that MRCA but people are telling me I need to have exact matches up to 67-111 y-STR and common SNPs which I think is unreasonable since FTDNA advertises those are to find MRCA under 8 generations away and the SNPs are not all mapped out for recent families how can I expect to find the same for ancient families? Family Finder also shows a very high >90% Orcadian / <10% Middle Eastern for nearly all participants that took Family Finder with shared cousin matches among all participants (even the other surname project where I suspect the ancient MRCA). Any thoughts?

  8. Pingback: What Does and Doesn’t a Y DNA Match Mean? | DNAeXplained – Genetic Genealogy

  9. Pingback: DNAeXplain Archives – Basic Education Articles | DNAeXplained – Genetic Genealogy

  10. Does the numbers below mean a Full, 1/2 or possible 3/4 Sibling.
    23andme had us listed as uncle/niece. Were 4 years apart. lol.

    gedmatch
    Largest segment = 157.0 cM
    Total of segments > 7 cM = 2,156.0 cM
    Estimated number of generations to MRCA = 1.4

    23andme
    34.1 shared, 57 segments

  11. I have compared my results and we have a MCRA of 3.8 generations. I believe my great-great-great grandfather is her great-great grandfather. Does that sound correct?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s