D. R. Schrider

IntroUNET: identifying introgressed alleles via semantic segmentation

This paper introduces IntroUNET, a deep learning method for identifying introgressed alleles using semantic segmentation, showing high accuracy in simulated and real data from Drosophila.

Tree sequences as a general-purpose tool for population genetic inference

The paper proposes a graph convolutional network to directly learn from tree sequences for population genetic inference, showing comparable or better performance than CNNs on benchmark tasks.

Allelic gene conversion softens selective sweeps

This study shows that gene conversion can create the appearance of soft sweeps even when selection acts on single-origin mutations, challenging previous assumptions about the prevalence of soft sweeps.

The genetic consequences of range expansion and its influence on diploidization in polyploids

This study uses simulations to show that autopolyploids face greater fitness costs during range expansion than diploids, while allopolyploids with disomic inheritance are shielded, and range expansion accelerates diploidization in autotetraploids.

This population does not exist: learning the distribution of evolutionary histories with generative adversarial networks

This paper trains a GAN to generate artificial population genetic alignments that mimic key features of real data across diverse evolutionary models.

Sex differences in inter-individual gene expression variability across human tissues

The paper identifies sex-differentially variable genes across human tissues and links them to biological functions and disease, suggesting a genetic basis and evolutionary model for sex-specific expression variability.

Timesweeper: Accurately Identifying Selective Sweeps Using Population Genomic Time Series

Timesweeper identifies selective sweeps in population genomic time-series data using a convolutional neural network, showing high accuracy and resolution in detecting selected variants and estimating selection coefficients.

Widespread introgression across a phylogeny of 155 Drosophila genomes

The paper analyzes 155 Drosophila genomes to detect widespread introgression across nine monophyletic radiations, revealing both ancient and recent hybridization events.

Deep ancestral introgression shapes evolutionary history of dragonflies and damselflies

The paper investigates introgression in dragonflies and damselflies, showing it is widespread and has shaped their evolutionary history, including deep ancestral introgression between superfamily lineages.

Discovery of ongoing selective sweeps within Anopheles mosquito populations using deep learning

This paper introduces a deep learning method to detect ongoing selective sweeps in mosquito populations, finding strong overrepresentation of sweeps at insecticide resistance loci.

A community-maintained standard library of population genetic models

The paper introduces stdpopsim, a community-maintained library for population genetic models, to standardize simulations and facilitate benchmarking of inference tools.

The Unreasonable Effectiveness of Convolutional Neural Networks in Population Genetic Inference

This paper applies convolutional neural networks to population genetic inference, showing they match or exceed current methods in accuracy and model selection.

diploS/HIC: an updated approach to classifying selective sweeps

The paper introduces diploS/HIC, a deep learning method that classifies selective sweeps using unphased genotypes, showing effectiveness at moderate sample sizes.

Machine Learning for Population Genetics: A New Paradigm

This paper reviews the application of supervised machine learning in population genetics, highlighting its potential to outperform traditional methods in analyzing genomic data.

Soft sweeps are the dominant mode of adaptation in the human genome

This study finds that soft sweeps, not hard sweeps, are the dominant mode of recent human adaptation across six populations.