🤖 AI Summary
This study addresses the low utilization efficiency of transcriptomic data and poor interpretability of knowledge transfer in cross-species phenotypic prediction. We propose an interpretable transfer learning framework grounded in orthologous gene relationships. Methodologically, we construct a bipartite graph modeling orthology between source and target species’ genes, and design a learnable species transformation layer constrained by the adjacency matrix mask, integrated with a pretrained feedforward network to enable directed mapping of gene expression spaces and downstream phenotypic prediction. Our key contribution is the first incorporation of bipartite graph structural priors as explicit weight constraints in neural networks—endowing the transformation layer with both transfer capability and functional orthology interpretability. Empirical evaluation on synthetic data demonstrates superior predictive performance and enhanced biological plausibility over baseline models. The framework establishes an interpretable, scalable computational paradigm for subsequent experimental validation.
📝 Abstract
We present Transfer Orthology Networks (TRON), a novel neural network architecture designed for cross-species transfer learning. TRON leverages orthologous relationships, represented as a bipartite graph between species, to guide knowledge transfer. Specifically, we prepend a learned species conversion layer, whose weights are masked by the biadjacency matrix of this bipartite graph, to a pre-trained feedforward neural network that predicts a phenotype from gene expression data in a source species. This allows for efficient transfer of knowledge to a target species by learning a linear transformation that maps gene expression from the source to the target species' gene space. The learned weights of this conversion layer offer a potential avenue for interpreting functional orthology, providing insights into how genes across species contribute to the phenotype of interest. TRON offers a biologically grounded and interpretable approach to cross-species transfer learning, paving the way for more effective utilization of available transcriptomic data. We are in the process of collecting cross-species transcriptomic/phenotypic data to gain experimental validation of the TRON architecture.