Transfer Orthology Networks

TL;DR

Transfer Orthology Networks (TRON) address cross-species transfer learning by exploiting orthologous relationships encoded as a bipartite graph $B$ between source and target species. The method masks a learnable conversion matrix $W_c$ with $B$ to form $x_t = (W_c ⊙ B) x_s$, and passes this through a pre-trained predictor $f$ to predict a phenotype, i.e., $y = f(x_t)$, with $f$ frozen during training. Only $W_c$ is updated, yielding a linear mapping from the source to target gene space guided by $B$. Orthology is defined via reciprocal best hits with $B_{ij}=1$ for RBH pairs, enabling interpretable mappings where $x_{t_i} = \sum_j (W_{c_{ij}} B_{ij}) x_{s_j}$ and larger $|W_{c_{ij}}|$ reflect stronger functional similarity across species. The paper also discusses a soft orthology constraint via regularization with hyperparameters $\alpha$ and $\beta$, and notes ongoing efforts to collect cross-species transcriptomic data for 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.

Figures (1)

• Figure 1: TRON Architecture