Transcriptome-wide association studies (TWAS) have proven successful in prioritizing genes and proteins whose genetically regulated expression modulates disease risk, but they ignore potential co-expression and interaction effects. Here we introduce the co-expression-wide association study (COWAS) method to identify pairs of co-expressed genes or proteins that are associated with complex traits. COWAS first trains models to predict co-expression conditional on genetic variation, and then tests for association between imputed co-expression and the trait while also accounting for direct effects from each exposure. We applied our method to plasma proteomic concentrations from the UK Biobank, identifying dozens of interacting protein pairs associated with cholesterol levels, Alzheimer's disease, and Parkinson's disease. Notably, our results demonstrate that co-expression between proteins may affect complex traits even if neither protein is detected to influence the trait when considered on its own.

The authors have declared no competing interest.

This work was supported by the National Institutes of Health (NIH) under grants R01 AG065636 and RF1 AG067924. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The authors also acknowledge the Minnesota Supercomputing Institute (MSI) at the University of Minnesota for providing high-performance computing resources that contributed to the research results reported within this paper.

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

Genotype, covariate, and protein expression data from the UK Biobank are available through the UK Biobank data access process (https://www.ukbiobank.ac.uk/enable-your-research). Access to the UK Biobank data was approved through UK Biobank Application #35107. Annotations for proteins assayed by the UK Biobank Pharma Proteomics Project are publicly available on Synapse (https://www.synapse.org/Synapse:syn51364943). Protein pairs with known interactions are publicly accessible in the HIPPIE web tool (https://cbdm-01.zdv.uni-mainz.de/~mschaefer/hippie). Publicly available GWAS summary statistics for cholesterol levels were downloaded from the Global Lipids Genetics Consortium website (https://csg.sph.umich.edu/willer/public/glgc-lipids2021). Publicly available GWAS summary statistics for Alzheimer's disease and Parkinson's disease were obtained from the NHGRI-EBI GWAS Catalog (https://www.ebi.ac.uk/gwas) under accession numbers GCST90027158 and GCST009325, respectively.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

Our software for COWAS is implemented in R and made available on GitHub under a GPL-3.0 open source license at https://github.com/mykmal/cowas. This GitHub repository also contains the scripts used for data quality control and batch processing. Fitted model weights for all protein expression and co-expression imputation models trained in this study are provided on Synapse at https://synapse.org/cowas.