Deep mendelian randomization: Investigating the causal knowledge of genomic deep learning models

Stephen Malina, Daniel Cizin and David A Knowles

PLOS Computational Biology, 2022, vol. 18, issue 10, 1-14

Abstract: Multi-task deep learning (DL) models can accurately predict diverse genomic marks from sequence, but whether these models learn the causal relationships between genomic marks is unknown. Here, we describe Deep Mendelian Randomization (DeepMR), a method for estimating causal relationships between genomic marks learned by genomic DL models. By combining Mendelian randomization with in silico mutagenesis, DeepMR obtains local (locus specific) and global estimates of (an assumed) linear causal relationship between marks. In a simulation designed to test recovery of pairwise causal relations between transcription factors (TFs), DeepMR gives accurate and unbiased estimates of the ‘true’ global causal effect, but its coverage decays in the presence of sequence-dependent confounding. We then apply DeepMR to examine the global relationships learned by a state-of-the-art DL model, BPNet, between TFs involved in reprogramming. DeepMR’s causal effect estimates validate previously hypothesized relationships between TFs and suggest new relationships for future investigation.Author summary: Chromatin marks such as transcription factor (TF) binding, accessibility, and histone modifications play a critical role in controlling cell behavior and identity. In recent years, multi-task deep learning (DL) models have achieved remarkable success at predicting these and other chromatin marks. However, it is unclear to what extent these models learn meaningful mechanistic, even causal, relationships between these variables. Our work aims to fill this gap by combining in silico mutagenesis, deep learning uncertainty estimation and causal inference (specifically Mendelian randomization, MR), into a framework we call DeepMR. We describe DeepMR, apply it to a simulation intended to test its ability to recover causal relationships between features from a learned model, and then use it to examine the relationships learned by a state-of-the-art DL model, BPNet. Our results suggest that DeepMR can estimate causal relationships under its stated assumptions and provide further evidence for previously hypothesized relationships between TFs identified by BPNet.

Date: 2022
References: View complete reference list from CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1009880 (text/html)
https://journals.plos.org/ploscompbiol/article/fil ... 09880&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pcbi00:1009880

DOI: 10.1371/journal.pcbi.1009880

Access Statistics for this article

More articles in PLOS Computational Biology from Public Library of Science
Bibliographic data for series maintained by ploscompbiol ().