Fig. 5: Association between increased ESC-mediated mutations and relapse.
From: Excised DNA circles from V(D)J recombination promote relapsed leukaemia
a, SVs per patient at diagnosis with a single cRSS (left) or two cRSSs (right) at the breakpoints plotted for patients who later relapsed (121 patients) versus those who remained in remission (29 patients). P value for the difference determined by a two-tailed Mann–Whitney U test. Single cRSS: P = 0.016; two cRSSs: P = 0.0006. SVs are from the TARGET study. Breakpoint junctions within antigen receptor loci are omitted. b, Right, pie chart showing all SVs at relapse with one cRSS at the breakpoint, two cRSSs or no cRSS for 83 matched patient samples; SVs at diagnosis were subtracted from those at relapse. Left, bar chart showing SVs at genes commonly mutated at relapse. P value for increased SVs involving single cRSSs at relapse-associated genes compared with all SVs involving a single cRSS calculated using a one-sided hypergeometric distribution; P = 0.0001. c, Frequency of breaks in 10 genes that most often acquire somatic mutations in relapsed BCP-ALL51 in the presence of the ESC (SJ) versus 12-RSS and 23-RSS controls (RSS) in a LAM-HTGTS experiment18. P value for co-localization of breaks detected by LAM-HTGTS and frequently mutated genes in relapsed ALL versus all genes mutated in ALL calculated using a one-sided hypergeometric distribution; P = 0.00077. n = 4. d, ddPCR of selected recombination junctions at diagnosis and relapse. Cell numbers normalized using GAPDH. Bars show mean values. n = 10 recombination junctions from four patients. e, ddPCR analysis of selected SJs at diagnosis and relapse for the indicated patients. Black dots show technical repeats. Bars show mean values. n = 10 SJs from 4 patients. f, Copy numbers of individual SJs at diagnosis for patients with ETV6–RUNX1 BCP-ALL (n = 16) who later relapsed or remained in remission. The horizontal line shows the highest normalized SJ copy number detected in healthy blood.
