On Tuesday, May 31, 2016 at 9:41:10 AM UTC-7, Andrew Lancaster via wrote: > taf wrote: > > "Sorry - that should have been that the 'steppe' ancestry is lower - > that something must have diluted it in England." > > Or perhaps the steppe component has been arriving more quickly on the > continent in recent millenia, than on the western edge such as the Isles. I must have been unclear. The Celtic parts of Britain (Scotland, Wales, Cornwall) have higher steppe ancestry than the Saxon part, and it was thought this was due to the disparity between the two populations. However, as I understand it, recent work with ancient Saxon DNA indicates that they have the same amount of steppe ancestry as the Celts, so while the English regions would be different that the Celtic regions with regard to population-specific markers, they should have similar proportions of steppe ancestry - if you mix pink and pink you don't get white. There must be some other contribution specifically to the population in the southeast of Britain and not in the west or north England, that is causing the steppe proportion to be lower. > I am not sure what samples are being used to represent the Celts and Old > Saxons but even in Britain itself That is an issue - Deinekes mentioned the paper and had a link so you can see for yourself. > And here is a quick prediction: very detailed Y DNA SNP testing is > probably the next big thing for genealogy as such (based on male lines > of course). The prices continue to drop but none of the big testing > companies are good at it yet, and to use it well will require clever web > interfaces. There is a reason for this. Autosomal SNPs spread throughout a population, then to all of mankind, due to recombination every generation. Hence the same set of SNP sites will be informative across all customers. mtDNA and Y-DNA SNPs are clonal - they are unique to the specific lineage in which they arose, so any given SNP locus is likely only to be informative for a small proportion of the customers - it takes sequencing the entire chromosome to find the individual informative ones for each lineage. Because the mt chromosome is so small, it is practicable to sequence it in its entirety for relatively low cost, but the Y chromosome basically requires whole-genome sequencing to find the Y-SNPs (you just throw out all of the non-Y sequence data). Thus my prediction - that targeted analyses will give way to whole-genome sequencing. A single analysis will tell you autosomal, mt and Y SNPs all at once and you can just pick out whichever parts of the data are relevant to the analysis you want to perform. Current cost is just over $1000 per genome, and dropping quickly. taf