Autoencoders for genomic variation analysis

Abstract

Modern biobanks are providing numerous high-resolution genomic sequences of diverse populations. In order to account for diverse and admixed populations, new algorithmic tools are needed in order to properly capture the genetic composition of populations. Here, we explore deep learning techniques, namely, variational autoencoders (VAEs), to process genomic data from a population perspective. We show the power of VAEs for a variety of tasks relating to the interpretation, compression, classification, and simulation of genomic data with several worldwide whole genome data sets from both humans and canids, and evaluate the performance of the proposed applications with and without ancestry conditioning. The unsupervised setting of autoencoders allows for the detection and learning of granular population structure and inferring of informative latent factors. The learned latent spaces of VAEs are able to capture and represent differentiated Gaussian-like clusters of samples with similar genetic composition on a fine scale from single nucleotide polymorphisms (SNPs), enabling applications in dimensionality reduction and data simulation. These individual genotype sequences can then be decomposed into latent representations and reconstruction errors (residuals), which provide a sparse representation useful for lossless compression. We show that different populations have differentiated compression ratios and classification accuracies. Additionally, we analyze the entropy of the SNP data, its effect on compression across populations, and its relation to historical migrations, and we show how to introduce autoencoders into existing compression pipelines.