Search

A growing body of evidence suggests that gene flow between closely related species is a widespread phenomenon. Alleles that introgress from one species into a close relative are typically neutral or deleterious, but sometimes confer a significant fitness advantage. Given the potential relevance to speciation and adaptation, numerous methods have therefore been devised to identify regions of the genome that have experienced introgression. Recently, supervised machine learning approaches have been shown to be highly effective for detecting introgression. One especially promising approach is to treat population genetic inference as an image classification problem, and feed an image representation of a population genetic alignment as input to a deep neural network that distinguishes among evolutionary models (i.e. introgression or no introgression). However, if we wish to investigate the full extent and fitness effects of introgression, merely identifying genomic regions in a population genetic alignment that harbor introgressed loci is insufficient-ideally we would be able to infer precisely which individuals have introgressed material and at which positions in the genome. Here we adapt a deep learning algorithm for semantic segmentation, the task of correctly identifying the type of object to which each individual pixel in an image belongs, to the task of identifying introgressed alleles. Our trained neural network is thus able to infer, for each individual in a two-population alignment, which of those individual's alleles were introgressed from the other population. We use simulated data to show that this approach is highly accurate, and that it can be readily extended to identify alleles that are introgressed from an unsampled "ghost" population, performing comparably to a supervised learning method tailored specifically to that task. Finally, we apply this method to data from Drosophila, showing that it is able to accurately recover introgressed haplotypes from real data. This analysis reveals that introgressed alleles are typically confined to lower frequencies within genic regions, suggestive of purifying selection, but are found at much higher frequencies in a region previously shown to be affected by adaptive introgression. Our method's success in recovering introgressed haplotypes in challenging real-world scenarios underscores the utility of deep learning approaches for making richer evolutionary inferences from genomic data.

Abstract The use of Bt proteins in crops has revolutionized insect pest management by offering effective season-long control. However, field-evolved resistance to Bt proteins threatens their utility and durability. A recent example is field-evolved resistance to Cry1Fa and Cry1A.105 in fall armyworm (Spodoptera frugiperda). This resistance has been detected in Puerto Rico, mainland USA, and Brazil. A S. frugiperda population with suspected resistance to Cry1Fa was sampled from a maize field in Puerto Rico and used to develop a resistant lab colony. The colony demonstrated resistance to Cry1Fa and partial cross-resistance to Cry1A.105 in diet bioassays. Using genetic crosses and proteomics, we show that this resistance is due to loss-of-function mutations in the ABCC2 gene. We characterize two novel mutant alleles from Puerto Rico. We also find that these alleles are absent in a broad screen of partially resistant Brazilian populations. These findings confirm that ABCC2 is a receptor for Cry1Fa and Cry1A.105 in S. frugiperda, and lay the groundwork for genetically enabled resistance management in this species, with the caution that there may be several distinct ABCC2 resistances alleles in nature.

The use of dsRNA to control insect pests via the RNA interference (RNAi) pathway is being explored by researchers globally. However, with every new class of insect control compounds, the evolution of insect resistance needs to be considered, and understanding resistance mechanisms is essential in designing durable technologies and effective resistance management strategies. To gain insight into insect resistance to dsRNA, a field screen with subsequent laboratory selection was used to establish a population of DvSnf7 dsRNA-resistant western corn rootworm, Diabrotica virgifera virgifera, a major maize insect pest. WCR resistant to ingested DvSnf7 dsRNA had impaired luminal uptake and resistance was not DvSnf7 dsRNA-specific, as indicated by cross resistance to all other dsRNAs tested. No resistance to the Bacillus thuringiensis Cry3Bb1 protein was observed. DvSnf7 dsRNA resistance was inherited recessively, located on a single locus, and autosomal. Together these findings will provide insights for dsRNA deployment for insect pest control.

Mimulus guttatus and M. nasutus are an evolutionary and ecological model sister species pair differentiated by ecology, mating system, and partial reproductive isolation. Despite extensive research on this system, the history of divergence and differentiation in this sister pair is unclear. We present and analyze a population genomic data set which shows that M. nasutus budded from a central Californian M. guttatus population within the last 200 to 500 thousand years. In this time, the M. nasutus genome has accrued genomic signatures of the transition to predominant selfing, including an elevated proportion of nonsynonymous variants, an accumulation of premature stop codons, and extended levels of linkage disequilibrium. Despite clear biological differentiation, we document genomic signatures of ongoing, bidirectional introgression. We observe a negative relationship between the recombination rate and divergence between M. nasutus and sympatric M. guttatus samples, suggesting that selection acts against M. nasutus ancestry in M. guttatus.

A central question in evolutionary biology concerns the developmental processes by which new phenotypes arise. An exceptional example of evolutionary innovation is the single-celled seed trichome in Gossypium ("cotton fiber"). We have used fiber development in Gossypium as a system to understand how morphology can rapidly evolve. Fiber has undergone considerable morphological changes between the short, tightly adherent fibers of G. longicalyx and the derived long, spinnable fibers of its closest relative, G. herbaceum, which facilitated cotton domestication. We conducted comparative gene expression profiling across a developmental time-course of fibers from G. longicalyx and G. herbaceum using microarrays with approximately 22,000 genes. Expression changes between stages were temporally protracted in G. herbaceum relative to G. longicalyx, reflecting a prolongation of the ancestral developmental program. Gene expression and GO analyses showed that many genes involved with stress responses were upregulated early in G. longicalyx fiber development. Several candidate genes upregulated in G. herbaceum have been implicated in regulating redox levels and cell elongation processes. Three genes previously shown to modulate hydrogen peroxide levels were consistently expressed in domesticated and wild cotton species with long fibers, but expression was not detected by quantitative real time-PCR in wild species with short fibers. Hydrogen peroxide is important for cell elongation, but at high concentrations it becomes toxic, activating stress processes that may lead to early onset of secondary cell wall synthesis and the end of cell elongation. These observations suggest that the evolution of long spinnable fibers in cotton was accompanied by novel expression of genes assisting in the regulation of reactive oxygen species levels. Our data suggest a model for the evolutionary origin of a novel morphology through differential gene regulation causing prolongation of an ancestral developmental program.