Fig. 6: IGK and IGL variants impact CDR3 physicochemical properties in the naïve Ab repertoire.

(A, B) For each CDR3 physicochemical property (x-axis), mean values were computed for each individual and tested for association (linear regression) with all common variants in IGK (A) and IGL (B). Barplots show (i) the number of QTL variants (Bonferroni-corrected) for each property, (ii) the −log10(P value) for lead variants, and (iii) the number of guQTL genes identified for the lead CDR3 property QTL variant. Summary statistics are provided in Supplementary Data 12. C Manhattan plot shows the −log10(P value) for all SNVs in the IGK locus tested for association with CDR3 aromaticity, with QTLs colored dark red and the lead QTL labeled. D Boxplot of the mean IGK CDR3 aromaticity with individuals separated by genotype at the lead QTL. E Boxplots of usages for seven IGK genes that are guQTLs at the lead CDR3 aromaticity variant (linear regression; P value < 3.7e−5). F BCR sequences that used the indicated V genes were selected from the Ab repertoire, then mean CDR3 aromaticity of each repertoire subset was computed and plotted with individuals separated by genotype at the lead CDR3 aromaticity QTL. Boxplots display the median, 25th percentile, 75th percentile, and whiskers that extend up to 1.5 times the inter-quartile range (IQR) from the respective percentiles. Data points outside the whiskers are also plotted.