Fig. 3: Condition-specific sQTL discovery in resting and activated chondrocytes.

a Venn diagram showing the overlap between sGenes identified in PBS and FN-f treated conditions. Linear mixed-effect modeling allowed the detection of 677 condition-specific sGenes (477 PBS-specific and 200 FN-f-specific). b The distribution of distances between sSNPs and the affected splice intron junctions (kb). c Percentage of sSNPs located within the same gene (within a gene) or outside the gene. d–f Examples of d shared (RHBDF2), e PBS-specific (SLC26A4), and f FN-f-specific (MAPK8) sQTLs. Each panel contains locus zoom plots showing the -log₁₀(p-value) of association between SNPs and splicing across genomic positions for both PBS and FN-f conditions. Gene structures are displayed beneath each locus zoom plot. Each panel also includes box plots displaying the relationship between genotype and inverse normal transformed PSI values for the indicated splice junction in both PBS and FN-f conditions (n = 101 independent donor samples per condition). Boxes indicate the interquartile range (25th–75th percentile) with a line at the median (50th percentile). Whiskers extend to the minimum and maximum values within 1.5× the interquartile range, with outliers shown as individual points. The figure shows the lead SNP ID, effect size (Beta), and p-value for each condition. Shared sQTLs show significant associations in both PBS and FN-f conditions. PBS-specific sQTLs show significant association only in PBS but not in FN-f condition. FN-f-specific sQTLs show significant association only in FN-f but not in PBS condition. Nominal p-values were calculated using QTLtools, which performs two-sided tests for association by converting correlation coefficients to p-values using a t-distribution of freedom (n-2, where n = 101 samples).