Heritability, pQTLs, and environmental influence on proteins involved in age, cardiovascular risk, and glucose tolerance using the SomaScan^® Assay

Background

Protein quantitative trait locus (“pQTL”) studies identify genetic variants that are statistically associated with protein levels. Results from the growing number of pQTL studies can be combined with genome-wide association studies to identify proteins that underlie the genetic risk of disease, thus revealing the mechanisms of disease and potential drug targets.

The SomaScan Assay is a highly multiplex protein assay that has facilitated the growth in pQTL studies. For affinity-based assays like SomaScan, cis-pQTLs provide evidence of specificity. However, a pQTL association will only be identified for a given protein if enough of the variation in the protein is due to genetics. Non-genetic sources of variation, such as differences in lifestyle or exposure to environmental risk factors, also influence the risk of disease and can be captured by proteomics.

Methods

• Compiled a list of cis-pQTLs from 15 studies run on the SomaScan Assay [1-15]
• Extracted heritability of aptamers from published results [14]. Calculated proportion of SomaScan Assay aptamers with cis-pQTLs or non-zero heritability (“genetic influence”)
• Compared to proportion of aptamers that are also strongly associated (FDR < 10%) with age, four-year cardiovascular disease risk (CVD), and oral glucose tolerance tests [16,17]
• Results shown here are based on the SomaScan v4 Assay, which measures ~ 5,000 proteins

Results

• Distribution of cis-SNP-based heritability stratified by ancestry for proteins in the SomaScan v4 Assay [14]. Proteins with non-significant (zero) heritability are shown in the left-most (gray) bar.
• Proportion of analytes with evidence of genetic basis (cis-pQTL or heritability > 0) stratified by respective endpoint association (FDR < 10%). Genetically influenced proteins are over-represented among endpoint associations, but up to 40% of associated, measured proteins do not have pQTLs. Endpoint associations, pQTLs, and heritability were identified using a 5K analyte assay, rather than the current 7K analyte assay.
• Regardless of pQTL status or heritability, each endpoint is associated (FDR < 10%) with a unique combination of analytes.

Conclusions

• Endpoint-associated analytes are more likely to have a cis-pQTL or have non-zero heritability than non-endpoint-associated analytes
• Most analytes are sensitive to environmental effects
  • Many endpoint-associated analytes do not have evidence of genetic influence
  • Heritability levels are generally modest
• Distinct sets of analytes respond to each endpoint
• The SomaScan Assay is highly multiplex and can identify hundreds to thousands of proteins associated with endpoints
  and with genetic variants

Authors

Brendan Epstein
Tina Lai
Ted Johnson
Michael A. Hinterberg
David P. Astling

SomaLogic Operating Co., Inc., Boulder, CO USA