ANDES: a novel best-match approach for enhancing gene set analysis in embedding spaces

Abstract

Embedding methods have emerged as a valuable class of approaches for distilling essential information from complex high-dimensional data into more accessible lower-dimensional spaces. Applications of embedding methods to biological data have demonstrated that gene embeddings can effectively capture physical, structural, and functional relationships between genes. This utility has primarily been demonstrated by using gene embeddings for downstream machine learning tasks. Much less has been done to examine the embeddings directly, let alone embeddings for a set of genes beyond calculating simple mean embeddings between sets. Here, we propose a novel best-match approach that considers gene similarity while reconciling gene set diversity. This intuitive method has important downstream implications for improving the utility of embedding spaces for various tasks. Specifically, we show how our method, combined with different gene embeddings encoding protein-protein interactions, can be used as a novel overrepresentation-based and rank-based gene set enrichment analysis method that achieves state-of-the-art performance. Additionally, when used with a multi-organism joint gene embedding, our method can facilitate functional knowledge transfer across organisms, allowing for phenotype mapping across model systems. Our flexible, straightforward best-match methodology can be extended to other embedding spaces where there is a diverse community structure between set elements.