Few questions

[ADormant]

@Schraiber
Hello, I've a few question about your method. Can it detect indirect ancestry and the amount and directionality of it? Your paper seem to suggest that mesolithic WHG, SHG, EHG Europeans aren't ancestral to modern Europeans but while Europeans certainly have limited genetic continuity other methods I've seen suggest modern Europeans do have small but notable indirect admixture from similar mesolithic/neolithic groups mediated mainly by Euro-HG admixed Middle Eastern groups.
Will your method be able to detect continuity and amount of ancestry from Paleolithic Europeans in WHG? Genetic relationship between paleolithic HGs such as Oase, Tianyuan, Ust'-Ishim, Kostenki, Vestonice, Goyet and El Miron and WHG seem to be still poorly understood.

[ADormant]

@Schraiber Europeans went through great bottleneck and very strong drift during the LGM which must have changed their allele frequencies. if look at ADMIXTURE results Paleolithic Europeans have very different genetic components than Mesolithic Ones and plot on PCAs nowhere near WHG. In fact they are ASI/South Asian shifted/admixed so this confounds the results. Yet some some methods such as QpAdm suggest that WHG indeed descended in at least in part from Vestonice and Goyet samples despite lacking that South Asian affinity.
Can your method recognise continuity despite very strong differential drift and/or mutation rate?

https://genetiker.wordpress.com/

http://www.open-genomes.org/analysis/PCA/Eurogenes_Ice_Age_Eurasians_PC_plot_1-2-3.html

https://pdfs.semanticscholar.org/5e94/f8c99ce3626e92be894f92ca604fb999130c.pdf

http://www.biorxiv.org/content/biorxiv/early/2016/05/25/033852.full.pdf

https://postimg.org/image/7ug497fc1/

[Schraiber]

Hi @ADormant, my apologies for an incredibly, incredibly delayed reply. I think the key thing to realize is that my method tests for a VERY strong definition of continuity, in which there is no outside gene flow between the ancient and modern samples. I'm currently thinking about how to use a similar approach to try to do something more like a "temporally aware Admixture" but it's still very much int he beginning stages. Perhaps an interesting place for you to look would be this paper: https://academic.oup.com/bioinformatics/article/33/14/2148/3002763