Rare Disease Research

GPX4 Interactome for Mechanistic Discovery and Drug Repurposing

Supporting material for an article submitted for consideration in Pacific Symposium on Biocomputing © 2027 World Scientific Publishing Co., Singapore, http://psb.stanford.edu/

This paper is written to support Little Champ Raghav's CureGPX4 Project

Novel (●) and known (●) interactors of GPX4

Key inferences for novel interactors of GPX4

SMDS (Spondylometaphyseal dysplasia, Sedaghatian type) is an ultra-rare developmental disorder caused by GPX4 mutations and characterized by skeletal, neurological, cardiac, and developmental abnormalities. The evidence below highlights how each predicted GPX4 interactor is linked to one or more of these biological processes or disease-relevant pathways.

These computationally predicted PPIs are supported by convergent network, phenotype, expression, and functional-module evidence; experimental validation remains necessary.

This work demonstrates the use of our web applications Wiki-Pi and LENS in rare disease research:
Wiki-Pi protein interaction web application
Wiki-Pi: Webserver of protein–protein interactions
LENS network and enrichment analysis web application
LENS: Enrichment and network studies of proteins

Kalyani B. Karunakaran, N. Balakrishnan, Madhavi K. Ganapathiraju*

Abstract: Rare diseases pose a major challenge for mechanistic and therapeutic discovery because of limited molecular data and small patient populations. We present an integrative network analysis framework for constructing disease-specific functional landscapes by combining protein-protein interactions (PPIs), disease-gene associations, phenotype annotations, tissue expression profiles, gene perturbation data, and drug-target relationships. We applied this framework to Spondylometaphyseal dysplasia, Sedaghatian type (SMDS), an ultra-rare lethal skeletal dysplasia caused by mutations in GPX4 gene. Disease-centric and gene-centric interaction networks identified enriched pathophenotypes, functional modules, tissue-specific expression patterns, and candidate therapeutic targets. The resulting functional landscape connected GPX4 with genes implicated in SMDS and related skeletal dysplasias, and with regulators of GPX4 expression, and was enriched for SMDS-relevant phenotypes, fetal-stage functional modules, and brain/testis-enriched genes. Seven high-confidence computationally predicted GPX4 interactors (APBA3, EGR4, FUT5, GAMT, GTF2F1, MATK, and ZNF197) emerged as high-priority candidates supported by convergent network, phenotype, and expression evidence. Integrative transcriptomic and network analyses prioritized eleven candidate repurposable drugs, with resveratrol additionally highlighted by network proximity. Nearly 75% of rare diseases are associated with single genes, several of which are understudied with limited experimental data. This study demonstrates a framework for studying a rare disease in the context of its protein interactome to generate mechanistic and therapeutic hypotheses.

*Corresponding author