Featured Posters: Through a Population Genetics Lens

View Poster Materials

Katharine Korunes1, Sandra Beleza2, Amy Goldberg1

1) Duke University, Durham, NC; 2) University of Leicester, Leicester, UK.

Recently admixed populations offer unique opportunities to study rapid evolution. Many standard population genetic methods do not work on the scale of tens of generations. Instead, we can leverage patterns in genetic ancestry in admixed populations to understand short-term evolution in general. On very short timescales, changes in allele frequencies may be difficult to observe and biased by factors like nonrandom mating. In contrast, changes in ancestry patterns in admixed populations can be observed within tens of generations. Using genome-wide SNP data from 564 individuals spanning 7 islands of Cape Verde, we infer the demographic and selective history of the last ~20 generations since founding. Cape Verdeans are admixed descendants of Portuguese colonizers and West African slaves who settled the islands during the 1400s. First, applying a model-based framework, we infer a positive correlation of inferred ancestry between mating pairs (r2 = 0.25-0.66). The strength of this relationship varies between islands, and this variation is consistent with historical patterns such as consanguinity on islands with stronger genetic evidence of assortative mating. Next, we show that runs-of-homozygosity (ROH) reflect the contributions of the source populations and patterns of assortative mating. Perhaps surprisingly, the admixed population has lower levels of overall ROH than African source populations. Breaking up ROH into 3 classes by length, we find that short and medium length ROH are often broken up by mismatched local ancestry decreasing genome-wide ROH levels. However, long ROH are actually enriched in the Cape Verdeans. These excess long ROH likely reflect small population sizes and mating preferences post admixture. ­Finally, we use patterns of local ancestry to test for selection. We identify a long African ancestry tract surrounding the Duffy-negative allele, which is protective against a malaria parasite. Next, we compared Santiago, the only island with endemic malaria, to the other Cape Verdean islands without malaria. Accounting for difference in admixture history, we find an excess of the protective allele on the island of Santiago, but not on other Cape Verdean islands. Together, these results provide insight into the population history of Cape Verde and show how admixed populations provide powerful test cases for understanding evolutionary processes within tens of generations.

View Poster Materials

Moein Rajaei1, Ayush S. Saxena1, Michael Snyder1, Robyn E. Tanny2, Erik C. Andersen2, Joanna Joyner-Matos3, Charles F. Baer1

1) University of Florida, Gainesville, FL USA; 2) Northwestern University, Evanston, IL USA; 3) Eastern Washington University, Cheney, WA USA.

The rate and spectrum of mutation are of fundamental importance in evolutionary biology.  Mutation accumulation (MA) experiments are the usual (and always the most efficient) way to estimate the properties of spontaneous mutation divorced from the influence of natural selection.  It is now apparent that the base-substitution spectrum of mutations accumulated in lab MA populations of C. elegans differ significantly and consistently from the standing site-frequency spectrum (SFS) in nature, with a greater proportion of transversions in the lab, especially C:G→A:T transversions.  There are two possible reasons: (1) natural selection skews the SFS away from the mutational spectrum, or (2) the spectrum of lab-accumulated mutations differs from that in nature.

One possible explanation for the discrepancy is that conditions in the lab result in increased oxidative damage to DNA relative to that in nature, perhaps associated with differences in metabolism (e.g., food ad libitum).  To test that hypothesis, we performed an MA experiment with a mutant strain of C. elegans, mev-1, that is known to experience elevated oxidative stress, resulting from a defective complex II of the mitochondrial electron transport chain.  If oxidative stress is a cause of the difference, we expect an even greater skew toward C:G→A:T transversions, which are a signature of oxidative damage to DNA.

Whole-genome sequencing of 24 mev-1 MA lines that had accumulated mutations for >100 generations revealed a rate and spectrum of base-substitution mutations that are essentially identical to other C. elegans MA lines.  Thus, there is no evidence that the discrepancy between the lab-accumulated and natural spectra is the result of increased oxidative damage under lab conditions.  Oxidative stress is known to increase the somatic mutation rate; apparently the germline is uniquely protected against oxidative damage.