Mo, Ziyi, Siepel, Adam (September 2023) Domain-adaptive neural networks improve supervised machine learning based on simulated population genetic data. bioRxiv. (Submitted)
PDF
2023_Mo_Domain-adaptive_neural_networks_improve_preprint.pdf - Submitted Version Available under License Creative Commons Attribution Non-commercial. Download (1MB) |
Abstract
Investigators have recently introduced powerful methods for population genetic inference that rely on supervised machine learning from simulated data. Despite their performance advantages, these methods can fail when the simulated training data does not adequately resemble data from the real world. Here, we show that this "simulation mis-specification" problem can be framed as a "domain adaptation" problem, where a model learned from one data distribution is applied to a dataset drawn from a different distribution. By applying an established domain-adaptation technique based on a gradient reversal layer (GRL), originally introduced for image classification, we show that the effects of simulation mis-specification can be substantially mitigated. We focus our analysis on two state-of-the-art deep-learning population genetic methods--SIA, which infers positive selection from features of the ancestral recombination graph (ARG), and ReLERNN, which infers recombination rates from genotype matrices. In the case of SIA, the domain adaptive framework also compensates for ARG inference error. Using the d omain- ada ptive SIA (dadaSIA) model, we estimate improved selection coefficients at selected loci in the 1000 Genomes CEU population. We anticipate that domain adaptation will prove to be widely applicable in the growing use of supervised machine learning in population genetics.
Item Type: | Paper |
---|---|
Subjects: | bioinformatics > computational biology > algorithms > machine learning organs, tissues, organelles, cell types and functions > tissues types and functions > neural networks bioinformatics > genomics and proteomics > genetics & nucleic acid processing > population genetics |
CSHL Authors: | |
Communities: | CSHL labs > Siepel lab |
SWORD Depositor: | CSHL Elements |
Depositing User: | CSHL Elements |
Date: | 6 September 2023 |
Date Deposited: | 29 Sep 2023 17:53 |
Last Modified: | 29 Sep 2023 17:53 |
PMCID: | PMC10002701 |
Related URLs: | |
URI: | https://repository.cshl.edu/id/eprint/41071 |
Actions (login required)
Administrator's edit/view item |