The impact of low-frequency genetic variants on serum protein levels
The mapping of protein quantitative trait loci (pQTLs) can provide molecular links between genotype and phenotype. Most such studies focus on common variants, but the effects of low-frequency (LF) variants remain underexplored. Focusing on cis-pQTLs, we integrated serum measurements of 7,596 proteins with genomic data, including LF variants (minor allele frequency [MAF] 0.1-1%), in 5,291 Icelanders to identify independent cis-pQTLs for 2,166 SOMAmers. Incorporating LF variants increased the number of detected genetic signals per protein, demonstrating widespread allelic heterogeneity in cis-acting regulation of serum proteins. LF pQTLs were enriched for coding variants in the respective protein-encoding gene, but also among distal secondary signals, revealing additional regulatory layers not captured by common variants alone. Proteins affected by common variant cis-pQTLs were more often secreted and exhibited tissue-specific expression, whereas proteins exclusively affected by LF variants were primarily from more constrained and biologically essential pathways. Expanding both protein coverage and the allele-frequency spectrum reveals a more complex and heterogeneous cis-regulatory architecture of circulating proteins.