Fig. 4: Longitudinal analyses and clonal dynamics during treatment.

a–d Leukemic burden (a) in ALL, days 15 and 29. b at the end-of-induction (EOI) in ALL (c), in AML, day 15 and after EO2I. d after EO2I for AML. e Linear diagram and fish plot showing clonal dynamics in P12. CR indicates complete remission by Fluorescence In Situ Hybridization (FISH) and hREL indicates hematologic relapse. Mutations were grouped into three clones and at CR, 1% of leukemia cells were detected. f–i Clonal evolution by fish plots, the treatment blocks are indicated and at the bottom, the day from diagnosis with circles, bone marrow (BM, grey), and peripheral blood (PB, red). Samples studied by targeted sequencing, WGS, and WES in bold, R=relapse. f An infant ALL that relapsed during maintenance. The major diagnostic clone contained deletions of PAX5, IKZF1, VPREB1 and an NRAS^G12D (green) and gained an ARID1A^S506C (red) and an NT5C2^R39Q in 27% of cells (yellow) at relapse, ALLR3 was started (without thiopurines), and the NTC52^R39Q-clone became undetectable. g For this infant, a PB-sample was taken 5 d before diagnosis; the purple clone became undetectable by day 19, and the leukemia cells increased to 60% 7 days later, CR was reached (by FISH), and the patient was transplanted. Around 3 months later, the diagnostic subclone (purple) that responded to induction therapy was detected (14%), and reached clonal dominance with no additional driver mutations. h A child where the KMT2A-r was detected at day 378 in 3% of cells, 92 days before relapse and 350 days after CR. The KRAS^G12N-relapse clone was detected at diagnosis (2%). i An AML where a subclonal NF1^R1947* outcompetes a larger PTPN11^D61V-clone at relapse. j The order of mutations at relapse as determined by single-cell sequencing. Mutations are gained sequentially with all relapse cells containing KRAS^G12N, PAX5::ZCCHZ7, CDKN1B^{D159_E1splice_region}, followed by expansion of cells with TP53, PTPN11, and CREBBP-mutations and a clone without additional known driver events.