bioMoR: Biology-Guided Mixture-of-Recursions for Effective Genomic Learning

2026-08-06

AI Summary

This study addresses the challenge that only a small subset of genes or pathways in high-dimensional omics data are discriminative, yet existing models treat all inputs uniformly, resulting in inefficiency and poor interpretability. To overcome this, the authors propose bioMoR, a novel framework that introduces mixture-of-recursives (MoR) mechanisms into gene- and pathway-level learning. bioMoR uniquely integrates biological priors at three key points within its MoR backbone: graph-structure-guided embedding refinement, structure-biased self-attention, and graph-aware routing. This design enables efficient allocation of computational resources and construction of interpretable representations. Evaluated across eight omics benchmarks, bioMoR achieves an average 8.2% improvement in macro F1 score, a 7.1% gain in balanced accuracy, 75% fewer parameters, and up to 58% reduction in FLOPs, while simultaneously identifying interpretable key genes and pathways.

Abstract

Transformer models for high-dimensional omics analysis process thousands of genes or pathways, although only a subset requires deep computation. Mixture-of-Recursions (MoR) improves efficiency through adaptive token-choice or expert-choice routing. We propose bioMoR, which, to the best of our knowledge, is the first framework to apply MoR to gene-level and pathway-level learning. Our contributions include identifying three locations for integrating structured biological knowledge within an MoR backbone: graph-based information sharing refines token embeddings, a structural bias guides self-attention toward biologically related tokens, and a graph-aware router uses neighborhood information to determine each token's recursion depth. These techniques are centered on our insight that additional knowledge of token interaction can effectively help models construct embeddings and select which tokens should be learned more deeply. Across eight benchmarks spanning diverse omics data types and evaluated under a unified five-fold cross-validation protocol, bioMoR improves average macro-F1 by 8.2 percentage points and balanced accuracy by 7.1 percentage points over the strongest biology-agnostic MoR baseline while using 75 percent fewer parameters and up to 58 percent fewer FLOPs than a non-recursive Transformer. The selected marker genes or pathways provide biological interpretability, while their token-specific recursion depths reveal how computation is allocated.

Problem

• omics analysis
• computational efficiency
• biological interpretability
• gene-level learning
• pathway-level learning

Innovation

• Mixture-of-Recursions
• biology-guided learning
• graph-aware routing
• genomic representation learning
• efficient Transformers