Friday, 22 April 2016

Big Y results - comparing TMRCA estimates

In the previous post, we looked in depth at the SNP markers identified by FTDNA and YFULL, and compared the reports from each company for similarities and differences. However, this post explores the topic of TMRCA (Time to Most Recent Common Ancestor) and thankfully this is a lot more straightforward.

TMRCA Estimates based on SNPs
Another useful piece of information from the YFULL analysis is the estimate for when the SNPs in this terminal block emerged and the TMRCA estimate between the two volunteers who have tested (TMRCA is Time to Most Recent Common Ancestor). We discussed in a previous post that the SNPs in this block emerged about 2200 years ago (or 200 BC) but today we are looking at the TMRCA between the two volunteers.

SNP emergence estimate & TMRCA estimate for GF1

Their TMRCA is estimated to be a mere 150 years before present (ybp) by which they mean 150 years prior to the approximate date of birth of these individuals, which (let's say) is approximately 1950. This gives a common ancestor born about the year 1800. However, the 95% Confidence Intervals around this estimate indicate that it could be anywhere from 75 years ago to 500 years ago. Or in other words, we can be 95% confident that the common ancestor was born some time between 1450 and 1875. This estimate could be refined if more people from Genetic Family 1 (GF1) were to do the Big Y test and upload their results to YFULL, but for now there is no pressing need to do so.

Calculation of the TMRCA estimate

TMRCA Estimates based on STRs
But how does this compare with TMRCA estimates based on STR markers? The TiP Report for the comparison of these two volunteers is detailed below.  It is based on a comparison of their STR markers at the 67-marker level. You can access your own TiP Report by clicking on the orange TiP icon beside each of your matches. This tells you how close or how distantly you are related (based on your STR values). You can select comparisons based on 12 markers, 25 markers, 37, 67 or 111 (depending on how many markers you have personally tested).

This analysis assesses the probability that the two individuals share a common ancestor on their direct male lines within the past "X" number of generations. This is a cumulative probability and so the probability increases over time and eventually reaches 100%.

TiP Report comparing Volunteer A (H1223)  with Volunteer B (164729)
(click to enlarge)
The 50% (midpoint) value is about 10 generations - in other words, there is a roughly 50% chance that the common ancestor was born within the last 10 generations, and a roughly 50% chance that it was sometime before that. The 5% and 95% probability levels are about 4 and 19 generations respectively. Allowing 30 years per generation, this gives us a midpoint TMRCA estimate of 300 years before present (ybp), with a 90% Confidence Interval of 120 to 570 years ago. And translating this into actual years gives us a midpoint estimate of 1650 (assuming an average year of birth for the two volunteers of about 1950), with a range of somewhere between 1380 to 1830 AD.

This TMRCA estimate based on STR values (1380-1650-1830) is not that close to the TMRCA estimate based on SNP values (1450-1800-1875). In fact, the midpoint estimate is out by 150 years.  Also, the range around the "best estimate" is very large, and could be quite far back in time (1380-1450). This is why we really have to be careful when interpreting TMRCA estimates - they may be out by several hundred years ... and in either direction!

However, there is an additional technique we can use to try to obtain more accurate assessments of  TMRCA estimates for the entire group, and that is something we will explore in a subsequent blog post.

Maurice Gleeson
April 2016

No comments:

Post a Comment