a. Number of somatic genetic alterations and epigenetic changes compared to normal counterparts along the course of the disease. Cases/time points with no grid lines correspond to unavailable data. b. Mutational burden, number of CNAs and number of SVs found in RT stratified according to the last therapy prior transformation. Targeted, targeted therapies. center line, median; box limits, upper/lower quartiles; whiskers, 1.5×interquartile range; points, individual samples. c. Copy number landscape of the studied cohort grouped by patient. The diagnosis, IGHV mutational status, last therapy prior RT, and total number of CNAs are indicated for each time point. d. Aggregated copy number profile of RT vs CLL. The first CLL samples (time point 1, T1) were considered. The plot shows the percentage of samples with gains (up) and losses (down). Among recurrent alterations found either in CLL or RT samples (n ≥ 5), deletions of 9p (PTPRD and CDKN2A/B) and deletions of 15q (MGA) were enriched in RT whereas deletions of ATM (11q), TP53 (17p), and 13q14 were found at similar frequencies in CLL and RT. e. Oncoprint of putative driver alterations. Samples, grouped by patient (patient id at the top), are represented by columns while genes in rows. Novel drivers in RT are labeled in blue. Genes are grouped according to their biological function or if they were previously described as potential driver genes in CLL and/or mature B cell lymphomas. Metadata including the type of therapy before RT, number of treatment lines before each sample, the spatial/longitudinal nature of the CLL/RT samples analyzed, IGHV mutational status, and diagnosis is detailed in the upper rows. In the main plot, mutations (SNVs and indels) are depicted with horizontal rectangles, CNAs using the background color of each cell, and SVs with vertical rectangles. The transparency of the color of mutations and CNAs indicates the cancer cell fraction (CCF). For patients lacking the germline sample (patient id indicated in gray), the CCF of the alterations could not be inferred and a CCF of 100% was used for illustrative purposes.