Seminars

February 22 2012

10:00 LSB 2320

This seminar is sponsored by Computational BioSciences Initiative and the Dept of EEB

Kirk Lohmueller
Department of Integrative Biology, University of California, Berkeley

Insights into human history and natural selection from genome sequencing data

Summary

Unlocking human evolutionary history is of great historical and anthropological interest, and is necessary for understanding the genetic basis of common disease. Next-generation sequencing technologies have led to an unprecedented amount of genetic variation data. Such data can be used to reconstruct human history and understand how natural selection operates in the human genome. Here I describe three projects that illustrate how this type of data can be combined with population genetic models to answer questions in human evolutionary genetics.

First, I use genetic variation data to reconstruct the history of Aboriginal Australians. We sequenced DNA from a 100-year old lock of hair from an Aboriginal Australian individual. Our analyses of this genome suggest that Aboriginal Australians derive from a separate and earlier wave of migration into East Asia than the migration that gave rise to mainland East Asian populations. This study supports the hypothesis that Aboriginal Australians represent one of the oldest human populations outside of Africa.

Second, I address the importance of natural selection in shaping genome-wide patterns of variation by characterizing patterns of allele frequency using low-coverage next-generation sequencing data. My analyses suggest that negative natural selection (i.e. the removal of deleterious mutations), rather than an abundance of sites under positive natural selection (i.e. adaptations), better explains several features of the data. This study supports the notion that the removal of weakly deleterious mutations by negative natural selection is a prominent feature across the genome, and that this process also affects nearby neutral sites.

Third, I evaluate how population demographic history influences the efficacy of negative selection in the coding regions of human genes. I show that European populations contain proportionally more weakly deleterious amino-acid changing polymorphisms than African American populations. Population genetic models suggest that this pattern can be explained by the differences in demographic history experienced by the two populations.

Hosted by Professor Bob Wayne

 

 

 

 

 

 



 

 

 

 

 

 

 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



this is idtest: