Extended Data Fig. 9: Increased mutations at relapse-associated genes correlate with high ESC levels.
From: Excised DNA circles from V(D)J recombination promote relapsed leukaemia
a, Characterisation of SVs with cRSS(s) at the breakpoint(s). Pie chart showing fraction of SVs that have two cRSSs at the breakpoints (consistent with cryptic recombination19,47) or one cRSS at the breakpoint, consistent with cut-and-run18, reintegration14,15,16 or RAG-mediated insertion49. Further breakpoint characteristics at single cRSSs (cut-and-run, reintegration or RAG-mediated insertion) were determined using a custom programme available from https://github.com/Boyes-Lab/Structural-Variants62. SVs from the TARGET study. n = 150 patients. b, RAG1 expression levels determined by RT-qPCR for BCP-ALL patients shown in Fig. 2a. Three patient groups were identified according to RAG1 expression and total normalised SJ reads: (i) High RAG1, high SJs (n = 14); (ii) High RAG1, low SJs (n = 15) and (iii) low RAG1, low SJs (n = 12). RAG expression data are normalised to HPRT expression levels. Data are presented as mean values ± s.d. The significance of the difference in expression levels was determined by a two-tailed Student’s t-test. 95% CI: 0.3122 to 1.360 (high RAG1 expression, high SJs versus low RAG1 expression, low SJs); 0.2294 to 0.9355 (high RAG1 expression, low SJs versus low RAG1 expression, low SJs) c, Total normalised SJ levels for the patient groups given in (b). Data are presented as mean values ± s.d. The significance of the differences was determined by a two-tailed Student’s t-test. 95% CI: 4.675 to 21.12 (high RAG1 expression, high SJs versus low RAG1 expression, low SJs); 5.431 to 19.94 (high RAG1 expression, high SJs versus high RAG1 expression, low SJs). d, SVs at genes that are commonly mutated in BCP-ALL or at relapse with a cRSS on one side of the breakpoint (blue), both sides (orange) or no cRSS (grey) for the patient groups given in (b) using WGS or WES compared to the corresponding germline sequence. The significance of increased SVs involving single cRSSs at relapse-associated or frequently mutated genes compared to all SVs involving a single cRSS was calculated using a one-sided hypergeometric distribution; P = 9.44×10−7 for the high RAG1, high SJ group (bottom panel). e, PCR amplification of the SYK gene using template DNA from patient samples taken at diagnosis or at relapse (indicated above the gel) as well as the BCP-ALL cell line, REH. R and NR indicate samples taken from patients who did, and who did not, later relapse. ddH2O indicates no template control. A 67 bp insertion in one allele of the SYK gene in patient R-8 increases between diagnosis and relapse. Quantification of the 272 and 339 bp bands is shown beneath the gel. For source gel data, see Supplementary Fig. 2b.
