TWAS Signature-Matching

Summary

Transcriptome-Wide Association Study (TWAS) signature-matching is an emerging in silico approach for genetically-informed drug prioritisation and repositioning. By integrating genome-wide association study (GWAS) summary statistics with tissue-specific expression quantitative trait loci (eQTL) datasets, this method imputes a genetically-predicted gene expression signature representing the causal changes associated with a disease. This signature is then matched against reference drug perturbation profiles (such as the Connectivity Map) to identify therapeutics whose expression signatures are strongly negatively correlated with the disease signature, indicating their potential to reverse disease-associated gene expression changes.

Framework and Parameter Optimization

While TWAS signature-matching provides a hypothesis-free drug repurposing tool that does not require prior knowledge of a drug's mechanism of action (MoA), its performance is highly sensitive to parameters evaluated in a systematic benchmark by the Shah Lab (as of 2026-07-19): - TWAS Tool: Utilizing sPrediXcan yields more robust drug enrichment compared to FUSION, as sPrediXcan avoids potential imputation errors of missing GWAS SNPs. - eQTL Tissue Selection: Biologically-relevant tissue-specific prediction models (e.g., GTEx liver models for LDL cholesterol) perform significantly better than multi-tissue models (like sMultiXcan) or whole blood models. - Drug Perturbation Cell Line: Selection of the cell line used for profiling drug signatures in databases is critical. Querying drug signatures in tissue-relevant lines (e.g., HEPG2 for LDL-C, HCC515 lymph node line for asthma) is required to successfully capture disease-drug relationships. - Similarity Metric: Spearman correlation is a more robust similarity metric for ranking compounds than the Normalized Connectivity Score (NCS), which treats up- and down-regulated genes separately. - Query Gene Set Size: Using small, unbalanced query gene sets (ranging from 5 up- and 5 down-regulated genes to 60 of each) rather than all statistically significant TWAS genes improves the signal-to-noise ratio by matching the scale of drug signatures (where median differential expression is ~120 genes).

Citations

• Chauquet, S., Jiang, J. C., Barker, L. F., Hunter, Z. L., Singh, G., Wray, N. R., McRae, A. F., & Shah, S. (2026). From GWAS to drug: A framework for drug candidate prioritisation using a gene expression signature matching approach. medRxiv. Source paper: 2026.04.22.26349470v2.full.pdf