Fig. 1: Relationship of mosaic chromosomal alterations (mCAs) with MBL.

Proportion of individuals with at least one canonical CLL-associated mCA (A) and autosomal mCA (B). Association of different categories of mCAs (defined below) with HC-MBL vs. controls in brown and LC-MBL vs. controls in dark gray (C) and HC-MBL vs. LC-MBL (D). ‘Controls’ refers to individuals in whom flow cytometry screening did not identify a B-cell clone in peripheral blood mononuclear cells. Covariates in (C) and (D) included age, sex, European ancestry, SNP-array type, source of DNA (whole blood or PBMC), and batch effects. del 17p is not shown in (C) as it was not found in any of the controls, and association statistics for del 11q, trisomy 12, and del 13q are not displayed for LC-MBL given their rarity in both this category and controls. Canonical CLL-associated mCA: del 6q, del 11q, trisomy 12, del 13q, del 17p, and copy-number neutral loss-of heterozygosity at 13q/ MIR16-1 at 13q/ MIR16-1. CLL-driver mCA: includes canonical CLL-associated mCAs as well as those that were among 179 candidate drivers of CLL identified in two large genomic studies of CLL [27, 28]. Lymphoid mCA: mCAs whose frequency was specifically enriched in individuals with lymphoid malignancies in comparison to myeloid malignancies [17]. Autosomal mCA without CLL-driver mCA or lymphoid mCA: autosomal mCAs in individuals who have neither a CLL-driver mCA nor a lymphoid mCA.