GenoHoption: Bridging Gene Network Graphs and Single-Cell Foundation Models

TL;DR

GenoHoption is a novel framework that integrates gene network graphs with foundation models for single-cell data, improving accuracy, reducing computation, and enabling few-shot learning.

Contribution

It introduces a method to incorporate gene network graphs into foundation models, balancing receptive field expansion and model efficiency.

Findings

Improves cell-type annotation accuracy by 1.27%

Enhances perturbation prediction by 3.86%

Reduces computational overhead significantly

Abstract

The remarkable success of foundation models has sparked growing interest in their application to single-cell biology. Models like Geneformer and scGPT promise to serve as versatile tools in this specialized field. However, representing a cell as a sequence of genes remains an open question since the order of genes is interchangeable. Injecting the gene network graph offers gene relative positions and compact data representation but poses a dilemma: limited receptive fields without in-layer message passing or parameter explosion with message passing in each layer. To pave the way forward, we propose GenoHoption, a new computational framework for single-cell sequencing data that effortlessly combines the strengths of these foundation models with explicit relationships in gene networks. We also introduce a constraint that lightens the model by focusing on learning the predefined graph structure while ensuring further hops are deducted to expand the receptive field. Empirical studies show that our model improves by an average of 1.27% on cell-type annotation and 3.86% on perturbation prediction. Furthermore, our method significantly decreases computational overhead and exhibits few-shot potential. GenoHoption can function as an efficient and expressive bridge, connecting existing single-cell foundation models to gene network graphs.