Prisoner Of Ice
04-14-2014, 03:39 AM
http://dienekes.blogspot.com/2014/04/neandertal-admixture-not-african.html
Although there has been much interest in estimating histories of divergence and admixture from genomic data, it has proved difficult to distinguish recent admixture from long-term structure in the ancestral population. Thus, recent genome-wide analyses based on summary statistics have sparked controversy about the possibility of interbreeding between Neandertals and modern humans in Eurasia. Here we derive the probability of full mutational configurations in nonrecombining sequence blocks under both admixture and ancestral structure scenarios. Dividing the genome into short blocks gives an efficient way to compute maximum-likelihood estimates of parameters. We apply this likelihood scheme to triplets of human and Neandertal genomes and compare the relative support for a model of admixture from Neandertals into Eurasian populations after their expansion out of Africa against a history of persistent structure in their common ancestral population in Africa. Our analysis allows us to conclusively reject a model of ancestral structure in Africa and instead reveals strong support for Neandertal admixture in Eurasia at a higher rate (3.4−7.3%) than suggested previously. Using analysis and simulations we show that our inference is more powerful than previous summary statistics and robust to realistic levels of recombination.
3.4−7.3%
There's now multiple studies saying it's about twice as high as initial estimates (other says up to 8%), but in every article etc. and from 23andme etc. they still go off the old values. Kind of annoying.
And this is the % of neanderthal specific genes not of ancestry - vast majority are actually shared with chimps, gorillas, etc. For comparison there's very, very few african specific ones, yet they still bang the drum for out of africa.
Although there has been much interest in estimating histories of divergence and admixture from genomic data, it has proved difficult to distinguish recent admixture from long-term structure in the ancestral population. Thus, recent genome-wide analyses based on summary statistics have sparked controversy about the possibility of interbreeding between Neandertals and modern humans in Eurasia. Here we derive the probability of full mutational configurations in nonrecombining sequence blocks under both admixture and ancestral structure scenarios. Dividing the genome into short blocks gives an efficient way to compute maximum-likelihood estimates of parameters. We apply this likelihood scheme to triplets of human and Neandertal genomes and compare the relative support for a model of admixture from Neandertals into Eurasian populations after their expansion out of Africa against a history of persistent structure in their common ancestral population in Africa. Our analysis allows us to conclusively reject a model of ancestral structure in Africa and instead reveals strong support for Neandertal admixture in Eurasia at a higher rate (3.4−7.3%) than suggested previously. Using analysis and simulations we show that our inference is more powerful than previous summary statistics and robust to realistic levels of recombination.
3.4−7.3%
There's now multiple studies saying it's about twice as high as initial estimates (other says up to 8%), but in every article etc. and from 23andme etc. they still go off the old values. Kind of annoying.
And this is the % of neanderthal specific genes not of ancestry - vast majority are actually shared with chimps, gorillas, etc. For comparison there's very, very few african specific ones, yet they still bang the drum for out of africa.