Interpretable Causal Representation Learning for Biological Data in the Pathway Space

TL;DR

This paper introduces SENA-discrepancy-VAE, a causal representation learning model that produces interpretable latent factors aligned with biological processes, enabling better understanding of gene and drug perturbations.

Contribution

The paper presents a novel CRL model that aligns latent factors with biological processes, improving interpretability without sacrificing predictive accuracy.

Findings

Achieves comparable predictive performance to non-interpretable models.

Infers biologically meaningful causal latent factors.

Provides an efficient encoder for biological process activity levels.

Abstract

Predicting the impact of genomic and drug perturbations in cellular function is crucial for understanding gene functions and drug effects, ultimately leading to improved therapies. To this end, Causal Representation Learning (CRL) constitutes one of the most promising approaches, as it aims to identify the latent factors that causally govern biological systems, thus facilitating the prediction of the effect of unseen perturbations. Yet, current CRL methods fail in reconciling their principled latent representations with known biological processes, leading to models that are not interpretable. To address this major issue, we present SENA-discrepancy-VAE, a model based on the recently proposed CRL method discrepancy-VAE, that produces representations where each latent factor can be interpreted as the (linear) combination of the activity of a (learned) set of biological processes. To this extent, we present an encoder, SENA-{\delta}, that efficiently compute and map biological processes' activity levels to the latent causal factors. We show that SENA-discrepancy-VAE achieves predictive performances on unseen combinations of interventions that are comparable with its original, non-interpretable counterpart, while inferring causal latent factors that are biologically meaningful.