Biological Random Walks: multi-omics integration for disease gene prioritization

TL;DR

This paper introduces Biological Random Walks (BRW), a novel method integrating multiple biological data sources for disease gene prioritization within the human interactome, demonstrating improved performance over existing methods.

Contribution

The paper presents a new multi-omics integration framework called BRW for disease gene prioritization, with extensive comparative evaluation showing its effectiveness.

Findings

BRW outperforms baseline methods in gene prioritization tasks

The framework effectively integrates diverse biological data sources

Code and datasets are publicly available for reproducibility

Abstract

Motivation: Over the past decade, network-based approaches have proven useful in identifying disease modules within the human interactome, often providing insights into key mechanisms and guiding the quest for therapeutic targets. This is all the more important, since experimental investigation of potential gene candidates is an expensive task, thus not always a feasible option. On the other hand, many sources of biological information exist beyond the interactome and an important research direction is the design of effective techniques for their integration. Results: In this work, we introduce the Biological Random Walks (BRW) approach for disease gene prioritization in the human interactome. The proposed framework leverages multiple biological sources within an integrated framework. We perform an extensive, comparative study of BRW's performance against well-established baselines. Availability and implementation: All code is publicly available and can be downloaded at \url{https://github.com/LeoM93/BiologicalRandomWalks}. We used publicly available datasets, details on their retrieval and preprocessing are provided in the supplementary material.