GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2009-07 > 1247384275


From: "Tim Janzen" <>
Subject: [DNA] supplemental TMRCA estimates of Y Adam and TMRCA estimates ofM168 and M9
Date: Sun, 12 Jul 2009 00:37:55 -0700


Dear All,
I decided to add some more options to my interclade TMRCA program
with the intent of seeing how TMRCA estimates are affected by clusters of
markers of assorted mutation rates. I added 5 new options as follows:
1. This option calculates the age in years for 67 marker haplotypes when
only including the 10 slowest mutating markers in the 67 marker FTDNA panel:
472, 578, 426, 455, 454, 641, 436, 490, 450, and 388.
2. This option calculates the age in years for 67 marker haplotypes when
only including the 10 medium slow mutating markers: 594, 640, 531, 617,
492, 392, 568, 590, 438, and 537.
3. This option calculates the age in years for 67 marker haplotypes when
only including the 10 medium mutating markers: 393, 565, 487, 437, 511,
448, 19, 406s1, 389i, and H4.
4. This option calculates the age in years for 67 marker haplotypes when
only including the 10 medium fast mutating markers: 572, 389ii, 520, 447,
391, 390, 557, 444, 442, and 446.
5. This option calculates the age in years for 67 marker haplotypes when
only including the 10 fastest mutating markers: 460, 607, 439, 481, 456,
570, 458, 534, 449 and 576.

To determine the effects of these marker selections on I decided to
use the new version of the program to rerun the TMRCA estimates on Y Adam
that I had done in March. See
http://archiver.rootsweb.ancestry.com/th/read/GENEALOGY-DNA/2009-03/12376891
68 and
http://archiver.rootsweb.ancestry.com/th/read/GENEALOGY-DNA/2009-07/12467369
45 for background. I ran all the new calculations using 9 67-marker
haplogroup A haplotypes, rather than the 7 I had available in March, but I
didn't add any new haplotypes to the datasets for the other haplogroups.
The haplotypes I used for the above calculations all came from their
respective FTDNA haplogroup projects. In most cases the samples in question
have been SNP tested by FTDNA but in some cases the haplotype was simply
predicted by FTDNA. I used 30 years as the generation interval. I continue
to use John Chandler's mutation rates with the exception of the YHRD option,
which uses the mutation rates for the 10 slowest mutating markers in the
YHRD as found at http://www.yhrd.org/Research/Loci. I generally excluded
all haplotypes that contained null values with the exception of some of the
haplogroups such as haplogroup E1b where exclusion of all of the haplotypes
that contained null values would have resulted in me having extremely few
haplotypes for testing. In such cases, I changed the marker value to one
below the marker value for the other haplotypes in that set.
I also ran TMRCA estimates for the node between haplogroup C and the
other haplogroups downstream of M168. I also ran TMRCA estimates for the
node between haplogroup L and the other haplogroups downstream of M9. I
also ran calculations for various nodes downstream from the NOP node.
The calculations are as follows:

Haplogroup TMRCA in years
Node 50 markers YHRD 24-slow slow medium-slow medium medium-fast fast
A/B node: 35089 8808 77394 147246 78635 8396 21794 13847
A/C node: 63495 18739 116242 304294 59912 23270 57241 19203
A/E node: 57130 13854 125948 361737 57458 16200 30445 10476
A/F node: 49312 13807 113808 321089 46486 20235 22255 6530
A/G node: 57469 12193 129701 317200 79527 20022 20878 15035
A/H node: 62347 14074 143096 398629 64638 19601 23412 11063
A/I node: 54443 11550 132353 351809 68182 18508 21409 11278
A/J node: 81744 12558 193603 535755 79287 12341 23725 17049
A/L node: 58009 33480 136617 333747 84388 15231 18448 22513
A/N node: 82927 38295 187740 361372 133969 45063 27563 17185
A/O node: 81833 31636 177607 467354 79509 20438 40433 15476
A/Q node: 61898 39224 139675 336568 88815 22948 23093 11306
A/R node: 63410 36125 146504 322738 111135 17916 24269 14764
A/T node: 71635 22171 163429 361335 117881 16835 26311 22927

Haplogroup TMRCA in years
Node 50 markers YHRD 24-slow slow medium-slow medium medium-fast fast
C/E node: 23215 15662 24630 64720 13315 18773 29727 15421
C/F node: 25223 19118 33021 49368 29350 17014 42137 12244
C/G node: 25708 19234 36668 66327 32070 19499 28333 10052
C/H node: 25876 9861 38941 63461 41554 15357 29585 17196
C/I node: 28799 20143 54122 153704 25871 22800 32188 5935
C/J node: 28033 15914 52793 141484 32407 17973 25033 8343
C/L node: 29439 31659 48638 60116 54036 21364 33659 12992
C/N node: 30541 32274 65567 61079 77120 24004 20129 5499
C/O node: 36653 26520 75484 188092 46906 18195 24475 14107
C/Q node: 27070 34492 47490 75663 55353 19490 25425 9271
C/R node: 23466 24452 42796 56407 54402 11323 24461 8428
C/T node: 23582 18928 41059 71283 47961 13141 15367 19035

Haplogroup TMRCA in years
Node 50 markers YHRD 24-slow slow medium-slow medium medium-fast fast
L/N node: 23201 19844 53776 17632 40674 43284 10442 9830
L/O node: 25328 15673 46078 135291 18998 18998 17471 11509
L/Q node: 18411 17508 31870 34417 30770 19458 19526 11894
L/R node: 16430 21078 33367 39263 34174 16755 12746 7670
L/T node: 17234 16215 27526 36107 24127 15147 19565 10143

Haplogroup TMRCA in years
Node 50 markers YHRD 24-slow slow medium-slow medium medium-fast fast
N/O node: 37415 18964 86112 156854 44186 43532 20209 10292
N/Q node: 22028 14484 42384 41622 38176 24989 17408 9509
N/R node: 17737 19312 44328 44273 39230 22768 10326 6693
Q/R node: 12533 14891 20590 23060 27635 11932 10725 7158
R1/R2 node: 14481 25532 29286 13017 44213 44213 10217 10774

The following are some observations following review of this data.
1. The TMRCA for Y Adam could be as much 300,000-400,000 years if the
calculations using the slowest markers are to be believed. I suspect that
the true age of Y Adam was much less than this, probably in the
80,000-150,000 year range. This would suggest that the mutation rate I am
using for one or more of the slowest mutating markers is lower than true
mutation rate for that marker or markers.
2. The TMRCA estimates when using just the fast and medium fast mutating
markers are clearly too low. This indicates that fast and medium fast
mutating markers should never be included in TMRCA calculations for older
haplogroups (any haplogroup suspected of being older than 20,000 years).
Medium mutating markers should be used with caution in TMRCA calculations
for older haplogroups, if at all.
3. Fast and medium fast mutating markers should probably be used with
caution in TMRCA calculations for younger haplogroups and subclades (those
2000-20,000 years old), if at all.
4. As one begins calculating TMRCA estimates on younger haplogroups and
subclades, there is increased volatility of the results when only the
slowest mutating markers are used.
5. Interestingly, the TMRCA estimates when using the medium-fast mutating
markers are in general higher than for those when using the medium mutating
markers when calculating the nodes for haplogroups A and C.
6. It is difficult to say which subset of markers produces the most
accurate TMRCA estimates for older haplogroups and subclades. It seems
clear that the faster mutating markers skew the TMRCA estimates lower than
the true TMRCA and thus must be omitted from the calculations. Probably a
group of 10-25 of the slowest mutating markers is all that can be safely
used. Even then the results are imprecise. I will probably try excluding
the results for the slowest mutating marker (DYS 472) and possibly also the
next slowest mutating markers (DYS 578 and 426) to see how the TMRCA
estimates for Y Adam and other haplogroups are affected.

Sincerely,
Tim Janzen


This thread: