Biology-inspired graph neural network encodes reactome and reveals biochemical reactions of disease

Functional heterogeneity of healthy human tissues complicates interpretation of molecular studies, impeding precision therapeutic target identification and treatment. Considering this, we generated a graph neural network with Reactome-based architecture and trained it using 9,115 samples from Genotype-Tissue Expression (GTEx). Our graph neural network (GNN) achieves adjusted Rand index (ARI) = 0.7909, while a Resnet18 control model achieves ARI = 0.7781, on 370 held-out healthy human tissue samples from The Cancer Genome Atlas (TCGA), despite the Resnet18 using over 600 times the parameters. Our GNN also succeeds in separating 83 healthy skin samples from 95 lesional psoriasis samples, revealing that upregulation of 26S-and NUB1-mediated degradation of NEDD8, UBD, and their conjugates is central to the largest perturbed reaction network component in psoriasis. We show that our results are not discoverable using traditional differential expression and hypergeometric pathway enrichment analyses yet are supported by separate human multi-omics and small-molecule mouse studies, suggesting future molecular disease studies may benefit from similar GNN analytical approaches.

Automated summary

Functional heterogeneity of healthy human tissues complicates molecular studies and precision therapeutic targets identification. A graph neural network (GNN) with Reactome-based architecture was trained using 9,115 samples from GTEx, achieving an adjusted Rand index (ARI) of 0.7909 on 370 healthy human tissue samples from TCGA. The GNN successfully separates healthy skin samples from lesional psoriasis samples, revealing a central mechanism in psoriasis. These results, supported by human multi-omics and mouse studies, suggest the potential benefit of GNN analytical approaches in future molecular disease studies.