Accurate identification of genotypes is critical in identifying de novo mutations, linking mutations with disease, and determining mutation rates. To call genotypes correctly from short-read data requires modeling read counts for each base. True heterozygotes may be affected by mapping reference bias and library preparation, leading to a distribution of reads that does not fit a 1:1 binomial distribution, and potentially resulting failure to call the alternate allele. Homozygous sites can be affected by the alignment of paralogous genes and sequencing error, which could incorrectly suggest heterozygousity. Previous work has modeled increased variance and skewed allele ratios to some degree. Here, we were able to model reads for all data as a mixture of Dirichlet multinomial distributions. This model has a better fit to the data than previously used models. In most cases we observed two distributions: one corresponds to a large proportion of heterozygous sites with a low reference bias and close-to-binomial distribution, and the other to a small proportion of sites with a high bias and overdispersion. The sites with high reference bias have not been previously identified as SNPs in extensive human genome research; thus, we believe these sites are not heterozygous in our data for the individuals studied here, and are falsely identified as heterozygous sites. We propose that this approach to modeling the distribution of NGS data provides a better fit to the data, which should lead to improved genotyping. Furthermore, the mixture of distributions may be used to suggest true and false positive de novo mutations. This approach provides an expected distribution of reads that can be incorporated into a model to estimate de novo mutations using reads across a pedigree.

Much work has been devoted to understanding the evolutionary processes shaping genetic variation across genomes. Studies have found that neutral polymorphism is reduced close to genes and in regions of low recombination, suggesting the effects of natural selection. However, the effect of selection on neutral sequence divergence between species remains ambiguous. While studies have reported correlations between divergence and recombination, theoretical arguments suggest selection may not affect divergence at linked neutral sites. Here we address these outstanding issues by examining how natural selection has affected divergence between distantly related species. We show that neutral divergence is negatively correlated with functional content and positively correlated with estimates of background selection from primates. These patterns persist even when comparing humans and mice, species that split 75 million years ago. Further, neutral divergence is positively correlated with recombination rate. The correlation increases when focusing on genic regions, and biased gene conversion cannot explain all of this correlation. These signatures suggest that natural selection has affected linked divergence between distantly related species. Coalescent models indicate that background selection can generate these patterns. Even when the contribution of ancestral polymorphism to divergence is small, background selection in the ancestral population can still explain a large proportion of the variance in divergence across the genome. Thus, the view that selection does not affect divergence at linked neutral sites needs to be reconsidered. Our work has important implications for understanding evolution of genomes and interpreting patterns of genetic variation.

Analysis of previously unpublished allele counts obtained from the French-San-Neanderthal-Chimpanzee alignment of the high quality DNA sequence of a Neanderthal from the Altai Mountains raises significant questions about the currently accepted phylogenetic model of the origins of Europeans. Previous estimates of the proportion of Neanderthal ancestry in present-day Europeans ranged between 1.3% and 2.7% supporting a recent Out-of Africa dispersal model followed by a low level of admixture with Neanderthals. However, analysis of the allele counts indicates the existence of an unidentified third archaic ancestor of Europeans, which diverged from its common ancestor with sub-Saharan Africans around 900 thousand years ago. This analysis shows that the relative proportions of derived alleles in the 0.0826% of the European genome that is not shared with the common ancestor of humans and chimpanzee are 13.6% Neanderthal, 32.3% sub-Saharan African and 54.2% third archaic ancestor. This analysis together with anthropological and archaeological evidence suggests a new model of human dispersal based on a Eurasian lineage in the Levant, which admixed with Neanderthals and descendants of African mtDNA haplogroup L3, followed by radiation from a basal admixed population around 55-50 Kya, with no subsequent major contribution to the European genome.

The genetic structure of human populations is often characterized by aggregating measures of ancestry across the autosomal chromosomes. While it may be reasonable to assume that population structure patterns are similar genome-wide in relatively homogeneous populations, this assumption may not be appropriate for admixed populations, such as Hispanics and African Americans, with recent ancestry from two or more continents. Recent studies have suggested that systematic ancestry differences can arise at genomic locations in admixed populations as a result of selection and non-random mating. Here, we propose a method, which we refer to as the chromosomal ancestry differences (CAnD) test, for detecting heterogeneity in population structure across the genome. CAnD uses local ancestry inferred from SNP genotype data to identify chromosomes harboring genomic regions with ancestry contributions that are significantly different than expected. In simulation studies with real genotype data from Phase III of the HapMap Project, we demonstrate the validity and power of CAnD. We apply CAnD to the HapMap Mexican American (MXL) and African American (ASW) population samples; in this analysis the software RFMix is used to infer local ancestry at genomic regions assuming admixing from Europeans, West Africans, and Native Americans. The CAnD test provides strong evidence of heterogeneity in population structure across the genome in the MXL sample ($p=4e-05$), which is largely driven by elevated Native American ancestry and deficit of European ancestry on the X chromosomes. Among the ASW, all chromosomes are largely African derived and no heterogeneity in population structure is detected in this sample.

Blindness has evolved repeatedly in cave-dwelling organisms, and investigating loss of sight presents an opportunity to understand the operation of fundamental evolutionary processes, including drift, selection, mutation, and migration. The observation of blind organisms has prompted many theories for their blindness, including loss-by-disuse and selection against eye development when eyes are not used. Here we have developed a model that shows just how strong selection must be for blind populations of a cave-dwelling species to evolve. We used approximations to determine levels of selection that would result in caves containing only sighted individuals, only blind individuals, or a stable population of both. We then incorporated drift into the model using simulations. Based on our model, strong selection is necessary for the evolution of blindness unless immigration rates are extremely low. Drift decreased the fixation of blindness in populations, although for intermediate levels of migration the level of selection required to fix blindness decreased substantially. We hypothesize that this degree of selection may be due to phototaxis in sighted individuals, who move toward the light leaving only blind individuals in the cave.

In clonally propagated crops, non-additive genetic effects can be effectively exploited by the identification of superior genetic individuals as varieties. Cassava (Manihot esculenta Crantz) is a clonally propagated staple food crop that feeds hundreds of millions. We quantified the amount and nature of non-additive genetic variation for key traits in a breeding population of cassava from sub-Saharan Africa using additive and non-additive genome-wide marker-based relationship matrices. We then assessed the accuracy of genomic prediction of additive compared to total (additive plus non-additive) genetic value. We confirmed previous findings based on diallel populations, that non-additive genetic variation is significant, especially for yield traits. Further, we show that we total genetic value correlated more strongly to observed phenotypes than did additive value, although this is constrained by low broad-sense heritability and is not beneficial for traits with already high heritability. We address the implication of these results for cassava breeding and put our work in the context of previous results in cassava, and other plant and animal species.

BACKGROUND: Recent advances in transcriptome sequencing have enabled the discovery of thousands of long non-coding RNAs (lncRNAs) across multitudes of species. Though several lncRNAs have been shown to play important roles in diverse biological processes, the functions and mechanisms of most lncRNAs remain unknown. Two significant obstacles lie between transcriptome sequencing and functional characterization of lncRNAs: 1) identifying truly noncoding genes from de novo reconstructed transcriptomes, and 2) prioritizing hundreds of resulting putative lncRNAs from each sample for downstream experimental interrogation. RESULTS: We present slncky, a computational lncRNA discovery tool that produces a high-quality set of lncRNAs from RNA-Sequencing data and further prioritizes lncRNAs by characterizing selective constraint as a proxy for function. Our filtering pipeline is comparable to manual curation efforts and more sensitive than previously published approaches. Further, we develop, for the first time, a sensitive alignment pipeline for aligning lncRNA loci and propose new evolutionary metrics relevant for both sequence and transcript evolution. Our analysis reveals that selection acts in several distinct patterns, and uncovers two notable classes of lncRNAs: one showing strong purifying selection at RNA sequence and another where constraint is restricted to the regulation but not the sequence of the transcript. CONCLUSION: Our novel comparative methods for lncRNAs reveals 233 constrained lncRNAs out of tens of thousands of currently annotated transcripts, which we believe should be prioritized for further interrogation. To aid in their analysis we provide the slncky Evolution Browser as a resource for experimentalists.

Custom sequence capture experiments are becoming an efficient approach for gathering large sets of orthologous markers with targeted levels of informativeness in non-model organisms. Transcriptome-based exon capture utilizes transcript sequences to design capture probes, often with the aid of a reference genome to identify intron-exon boundaries and exclude shorter exons (< 200 bp). Here, we test an alternative approach that directly uses transcript sequences for probe design, which are often composed of multiple exons of varying lengths. Based on a selection of 1,260 orthologous transcripts, we conducted sequence captures across multiple phylogenetic scales for frogs, including species up to ~100 million years divergent from the focal group. After several conservative filtering steps, we recovered a large phylogenomic data set consisting of sequence alignments for 1,047 of the 1,260 transcriptome-based loci (~630,000 bp) and a large quantity of highly variable regions flanking the exons in transcripts (~70,000 bp). We recovered high numbers of both shorter (< 100 bp) and longer exons (> 200 bp), with no major reduction in coverage towards the ends of exons. We observed significant differences in the performance of blocking oligos for target enrichment and non-target depletion during captures, and observed differences in PCR duplication rates that can be attributed to the number of individuals pooled for capture reactions. We explicitly tested the effects of phylogenetic distance on capture sensitivity, specificity, and missing data, and provide a baseline estimate of expectations for these metrics based on nuclear pairwise differences among samples. We provide recommendations for transcriptome-based exon capture design based on our results, and describe multiple pipelines for data assembly and analysis.

Whole-genome low-coverage sequencing has been combined with linkage-disequilibrium (LD) based genotype refinement to accurately and cost-effectively infer genotypes in large cohorts of individuals. Most genotype refinement methods are based on hidden Markov models, which are accurate but computationally expensive. We introduce an algorithm that models LD using a simple multivariate Gaussian distribution. The key feature of our algorithm is its speed, it is hundreds of times faster than other methods on the same data set and its scaling behaviour is linear in the number of samples. We demonstrate the performance of the method on both low-coverage and high-coverage samples.