Prioritizing disease genes with an improved dual label propagation framework

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Disease gene prioritization aims to identify genes responsible for specific phenotypes, crucial for understanding inherited diseases and advancing drug development.
• Existing network-based methods often struggle with false positive protein-protein interactions (PPIs), potentially leading to inaccurate disease gene identification.
• Previous approaches utilized basic label propagation on fixed PPI networks, failing to account for data noise and its impact on results.

Purpose of the Study:

• To propose a novel network-based framework, IDLP, for prioritizing candidate disease genes.
• To address the challenge of false positive protein-protein interactions in disease gene prioritization.
• To develop a robust method for identifying disease genes, especially for phenotypes with limited known genetic associations.

Main Methods:

• IDLP employs a label propagation framework across both PPI and phenotype similarity networks.
• It models and corrects for biases introduced by false positive protein interactions by learning the PPI and phenotype similarity matrices.
• The method refines noisy training matrices to enhance the accuracy of disease gene prioritization.

Main Results:

• IDLP demonstrated superior performance compared to eight state-of-the-art methods on OMIM datasets.
• The framework effectively prioritizes candidate disease genes even when few disease genes are initially known.
• Literature verification confirmed the high accuracy of IDLP in identifying novel disease-associated genes, highlighting its potential for biological discovery.

Conclusions:

• The IDLP model offers an effective solution for disease gene prioritization, particularly for understudied phenotypes.
• It provides a powerful tool for biologists to discover new disease genes, enhancing our understanding of genetic disorders.
• IDLP's ability to handle noisy data and limited prior knowledge makes it valuable for advancing genetic research.