bioMoR: Integrating Biological Knowledge into Mixture-of-Recursions for Genomics

bioMoR applies Mixture-of-Recursions to genomic learning by integrating structured biological knowledge through graph-based embedding refinement and a graph-aware router. This approach uses neighborhood information to adaptively determine token recursion depth for gene-level and pathway-level analysis.