GRAPE: Heterogeneous Graph Representation Learning for Genetic Perturbation with Coding and Non-Coding Biotype

2025-05-06

AI Summary

Existing gene regulatory network (GRN) inference methods overlook functional heterogeneity among genes—such as coding versus non-coding roles—and rely solely on coarse-grained metrics, limiting their ability to capture biologically realistic regulatory mechanisms. To address this, we propose BioHGNN, a biotype-aware heterogeneous graph neural network that explicitly incorporates gene biotype information into genetic perturbation effect prediction for the first time. BioHGNN initializes node representations using dual-modality features: DNA sequence embeddings from the Nucleotide Transformer and semantic embeddings of gene descriptions derived from a large language model. It further employs dynamic graph structure learning to refine the heterogeneous GRN topology. This design enables functionally specific modeling and joint learning of implicit regulatory interactions. Evaluated on multiple public perturbation datasets, BioHGNN achieves state-of-the-art performance, significantly improving both perturbation effect prediction accuracy and GRN reconstruction fidelity.

Technology Category

• Machine Learning: Bio-inspired Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

Predicting genetic perturbations enables the identification of potentially crucial genes prior to wet-lab experiments, significantly improving overall experimental efficiency. Since genes are the foundation of cellular life, building gene regulatory networks (GRN) is essential to understand and predict the effects of genetic perturbations. However, current methods fail to fully leverage gene-related information, and solely rely on simple evaluation metrics to construct coarse-grained GRN. More importantly, they ignore functional differences between biotypes, limiting the ability to capture potential gene interactions. In this work, we leverage pre-trained large language model and DNA sequence model to extract features from gene descriptions and DNA sequence data, respectively, which serve as the initialization for gene representations. Additionally, we introduce gene biotype information for the first time in genetic perturbation, simulating the distinct roles of genes with different biotypes in regulating cellular processes, while capturing implicit gene relationships through graph structure learning (GSL). We propose GRAPE, a heterogeneous graph neural network (HGNN) that leverages gene representations initialized with features from descriptions and sequences, models the distinct roles of genes with different biotypes, and dynamically refines the GRN through GSL. The results on publicly available datasets show that our method achieves state-of-the-art performance.

Problem

• Predict genetic perturbations to identify crucial genes efficiently
• Improve gene regulatory networks by leveraging gene-related information
• Model distinct roles of genes with different biotypes

Innovation

• Uses pre-trained models for gene feature extraction
• Incorporates gene biotype information for regulation
• Employs heterogeneous graph neural network (HGNN)