Fig. 1: MPRA identified 39 functional variants from 16 melanoma GWAS loci.

a MPRA workflow. Oligo libraries were synthesized using 145 bp of sequence encompassing each variant with risk or protective alleles or a scrambled sequence for core 21 bases in both forward and reverse directions, that was flanked by enzyme recognition sites and sequencing primer sequences, as well as 10 bp barcodes (10 tags per unique sequence). Libraries were cloned into luciferase constructs with or without a minimal TATA promoter. Cloned libraries were then transfected into HEK293FT cells or UACC903 melanoma cells to generate expressed RNA tag libraries. Both input DNA and RNA libraries were sequenced to assess the tag counts associated with the test sequences. Luc: luciferase gene, AAAAA: poly-A tail. b Volcano plots of MPRA results for each melanoma GWAS locus. Inverse P-values and effect sizes of allelic difference from UACC903 transfections are shown for each of the 16 loci tested. A two-sided Wald test with robust sandwich type variance estimate was used. Multiple comparisons were adjusted using the Benjamini and Hochberg method. Dashed horizontal lines indicate the FDR 1% cutoff for allelic difference in the UACC903 set. The most significant variant from each locus is labeled. Putative function of 39 significant MPRA variants are shown as activator (red circle), repressor (blue circle), or both (purple circle) (expression levels of either allele is higher, lower, or higher and lower than those of scrambled sequence, respectively). Gray variants above the FDR 1% cutoff are those that failed additional criteria (allelic difference in the combined data or significant departure from the scrambled control). No significant variants were identified from the loci on Chr6p22.3 and Chr7p21.1. Source data are provided as a Source Data file.