It was brought to my attention last week, thanks to a comment on this blog made by the user 'Umi', that another thesis on East African DNA variation was publicly available online:
Complex Genetic History of East African Human Populations
This is also an extensive thesis with a wealth of information akin to Plaster's thesis, the primary differences being that this one was more focused on parts of East Africa that are found further to the South of Ethiopia, and in addition to uni-parental analysis, it also included some Autosomal model-based inference, albeit of quite low resolution in today's standards; 848 microsattelites and 479 indels (refer to Tishkoff et al. 2009 for marker details).
Due to the extensive nature of the report I haven't had a chance to cover its entire scope, instead, for starters, I have first focused on the YDNA data by creating a relative frequency chart from the results reported in Fig. 3.3.2.
Several things to initially point out here,
- The report outlines the discovery of 4 new SNPs, TL1-4. The first two were found in Haplogroup B and downstream from B-M150 and B-M112 respectively. The last two, TL3 and TL4, were found in haplogroup E and downstream from E-U174 and E-V32 respectively. Incidentally, the fourth SNP that is under E-V32, TL4, could potentially be the same as Z808/Z809 as identified recently by the geneological community, however, as the report does not give the Y-Chromosome location of the SNP in a NCBI Build 36/37 format, this can not be verified, at least by me, at the moment.
- A couple of the frequency results in Fig. 3.3.2 do not add up, in particular, the frequency results for the Boni and the Baggara, but also to a lesser extent for the Kanuri and Teita. I have labeled the missing frequency results with a “?” in the relative charts for those specific populations.
- The Burji and Konso are labeled as being only from Kenya throughout the report, however most Burji are from Ethiopia, and the Konso are exclusively found in Ethiopia, I have reflected this in the charts.
- STR data is not readily available to perform TMRCA estimates on, however, some TMRCA results are reported using Zhivotovsky's rates in Table 3.3.1, nevertheless, these are estimates only for different lineages found in the dataset for all the samples and not necessarily comparing TMRCAs in the different populations under study.
- J-M62, while a subclade of J-M267, is not the main subclade of J-M267 found in East Africa, that would be J-P58, therefore, the results for J-12f2.1 (x M62, M172) reported, may after all be, or largely include, J-P58 lineages, off-course those results could also include variants of J-M267 other than J-P58 and J-M62 as well since the SNP was not directly tested.
- E-P2* lineages are abundantly found (> 30%) in the Konso, Burji and Mbugwe, however on closer examination and correlation with current data, these could be E-M329, E-V38* or even E-M215*, as none of these SNPs were directly tested. Genuine E-P2* lineages would be positive for E-P2 and negative for V38 and M215 (See Trombetta et al. 2011)
- Similarly, the E-M35* lineages reported could be members of relatively newly discovered lineages of E-Z830*( See this post for details), or some of the untested variantes of E-M35, i.e. E-V42, V92 and maybe even E-V68 (x M78)
The study is absolutely great: it deals with almost every single aspect of the highly complex African genetic landscape and does so in a very comprehensive and well illustrated manner. I have just browsed it over as of now but there's already an issue that is bugging me: does the author propose to revise mtDNA phylogeny linking L0 and L1 in a single haplogroup (as used to be believed in the past) or is just an artifact of his academic methodology ("Bayesian method as designed by Mr. Bayes") and therefore I just should ignore that detail altogether? But otherwise I'm just loving it...
ReplyDeleteI haven't really looked at the mtDNA portion of the paper yet
DeleteThis comment has been removed by the author.
ReplyDelete