Fig. 2: A catalogue of CSR junction characteristics to assess immunocompetence and DNA repair in human and mouse.

a Overview of n = 16 representative of n = 68 total SWIBRID features, grouped into 5 indicated categories. Small schemes were used to depict the respective features. Breakpoint matrix features include a histogram of breakpoints occurring between the donor S region (x-axis) and the acceptor S region (y-axis) with rectangles and darkness denoting breaks and their frequency, respectively. b Percentage of variance explained by donor identity, batch, or sequencing method for selected features from each category. c Box plots of explained variance for all features. “Robust” features with > 30% variance explained by donor, <20% by sequencing, and <25% by batch were selected for downstream analyses. d Range of selected features in two different cohorts of healthy donors (C1: 98 samples of 40 donors; C2: 41 samples of 14 donors). e Left: break_dispersion_SG1 values (female, n = 20; male, n = 20). Right: occupancy_top_clone values (age < 50, n = 21; age ≥ 50, n = 19). P-value from two-sided Wilcoxon test (*: p < 0.05, ***: p < 0.001). f Box plots of representative robust features generated with splenocytes and lymph node cells of immunized mice (grey, n = 12), combined with a re-analysis of CSR long-read sequencing data from Vincendeau et al.³⁰ (violet, n = 3). g PCA of re-analyzed Vincendeau data³⁰ using 32 of 44 robust features, excluding features related to isotypes missing in the dataset and features missing in mouse compared to human. h Comparison of scaled mean features of genotypes (Vincendeau data). Significantly different features are indicated using larger symbols (adj. p-value < 0.05 from two-sided t-test, with Benjamini–Hochberg correction). Boxes in (d and f) indicate 25th to 75th percentile; whiskers extend to largest/smallest value no further than 1.5× interquartile range, lines indicate median.