POEMS: Product of Experts for Interpretable Multi-omic Integration using Sparse Decoding
This work addresses the challenge of integrating multi-omic data for disease analysis, offering a method that balances interpretability and predictive accuracy, which is incremental as it builds on existing deep generative models by enhancing interpretability without linearization.
The paper tackles the trade-off between predictive performance and interpretability in deep generative models for multi-omic data integration by introducing POEMS, an unsupervised probabilistic framework that achieves competitive clustering and classification performance in a cancer subtyping case study while providing interpretability through sparse connections and a gating network.
Integrating different molecular layers, i.e., multiomics data, is crucial for unraveling the complexity of diseases; yet, most deep generative models either prioritize predictive performance at the expense of interpretability or enforce interpretability by linearizing the decoder, thereby weakening the network's nonlinear expressiveness. To overcome this tradeoff, we introduce POEMS: Product Of Experts for Interpretable Multiomics Integration using Sparse Decoding, an unsupervised probabilistic framework that preserves predictive performance while providing interpretability. POEMS provides interpretability without linearizing any part of the network by 1) mapping features to latent factors using sparse connections, which directly translates to biomarker discovery, 2) allowing for cross-omic associations through a shared latent space using product of experts model, and 3) reporting contributions of each omic by a gating network that adaptively computes their influence in the representation learning. Additionally, we present an efficient sparse decoder. In a cancer subtyping case study, POEMS achieves competitive clustering and classification performance while offering our novel set of interpretations, demonstrating that biomarker based insight and predictive accuracy can coexist in multiomics representation learning.