Sample collection, library preparation and analysis of WES data

Genomic DNA was extracted from cryopreserved tumour specimens at diagnosis and from matched normal PBMCs at complete remission, followed by WES library preparation using SureSelectXT Target Enrichment System Kit (Agilent Technologies, Santa Clara, CA, USA) and SureSelectXT Human All Exon v5 Capture Library Kit (Agilent Technologies) for Illumina Multiplexed Sequencing. The sequencing was performed with the HiScanSQ (Illumina, San Diego, CA, USA) using the 100-bp paired-end method. Variant detection was performed by three programs with different algorithms: MuTect (version 1.1.4), VarScan2 (version 2.3.8), and Strelka (version 1.0.14)^40–42. Among the candidate variants extracted by those algorithms, we considered a true variant when at least two of the three programs defined it as a positive variant. To determine common mutational signatures, we used an R package pmsignature (version 0.2.1) with default parameters⁴³. To define enriched mutational motifs in the whole exome, immunoglobulin locus, and BCL2 locus, a Fisher exact test was used to compare the expected frequency of each motif with the observed mutation rate. A Wilcoxon rank sum test was used to compare numbers of mutations between the groups. A P-value of less than 0.05 was considered significant. See the Supplementary Methods for additional details of next generation sequencing library preparation, sequencing and analysing data.