Fig. 2: Overarching analysis of screening data.

We performed MPRA in the MM plasma cell lines L363 and MOLP8. a Variant log₂ scores for the L363 and MOLP8 screens. For each variant, log₂ reflects the transcriptional activity of the alternative relative to the reference allele. Scores were calculated based on barcode activity estimates in all six genomic contexts (i.e., across both strands and all three sliding windows) and three replicates per cell line. Variants with strong effects (absolute log₂ score >0.2) in either screen are indicated in red. Pearson r and two-sided P values are shown. b Calculating log₂ scores using either positive-strand (x-axis) or negative-strand constructs (y-axis) for the same variant, we did not observe strand bias. c As an additional assay validation step, we carried out luciferase experiments for 20 variants, showing a significant positive correlation between the MPRA effect (x-axis) and the luciferase effect (y-axis). d g-chromVAR analysis of screened variants, weighted by their L363 log₂ scores showed enrichment of variants with strong MPRA scores in genomic regions with accessible chromatin in plasma cells, consistent with our assay selecting variants with endogenous regulatory activity. e Numbers of significant variants with FDR <5% in the two cell lines. f Numbers of variants showing both FDR <5% and strong effects (absolute log₂ score >0.2).