53The landscape of structural variants (SVs) in multiple myeloma remains poorly 54 understood. Here, we performed comprehensive classification and analysis of SVs in 55 multiple myeloma, interrogating a large cohort of 762 patients with whole genome and 56 RNA sequencing. We identified 100 SV hotspots involving 31 new candidate driver 57 genes, including drug targets BCMA (TNFRSF17) and SLAMF7. Complex SVs, 58 including chromothripsis and templated insertions, were present in 61 % of patients and 59 frequently resulted in the simultaneous acquisition of multiple drivers. After accounting 60 for all recurrent events, 63 % of SVs remained unexplained. Intriguingly, these rare SVs 61were associated with up to 7-fold enrichment for outlier gene expression, indicating that 62 many rare driver SVs remain unrecognized and are likely important in the biology of 63 individual tumors. 64 65 whole chromosome gains later in tumor evolution 13,20 . The second category is made up of 89 all other recurrent copy number alterations (CNAs), one or more of which are present in 90 virtually all patients 19,20 . Among these, gain of chromosome 1q21 often occurs early in 91 tumor evolution, while loss of tumor suppressor genes tends occur later 20-23 . The 92 structural basis of these established drivers remains poorly understood, and there is a lack 93 of data about the role of SVs beyond the most recurrent aberrations. 94We recently reported the first comprehensive SV characterization using WGS of 95 sequential samples from 30 multiple myeloma patients 20 . Despite the small sample set, 96SVs emerged as key drivers during all evolutionary phases of disease, and we identified a 97 high prevalence of three main classes of complex SVs: chromothripsis, templated 98 insertions and chromoplexy 20 . In chromothripsis, chromosomal shattering and random 99 rejoining results in a pattern of tens to hundreds of breakpoints with oscillating copy 100 number across one or more chromosomes 24 . Templated insertions are characterized by 101 focal gains bounded by translocations, resulting in concatenation of amplified segments 102 from two or more chromosomes into a continuous stretch of DNA, which is inserted back 103 into any of the involved chromosomes 2,20 . Chromoplexy similarly connects segments 104 from multiple chromosomes, but the local footprint is characterized by copy number 105 loss 25 . Importantly, complex SVs represent large-scale genomic alterations acquired by 106 the cancer cell at a single point in time, potentially driving subsequent tumor evolution. 107Here, we present the first comprehensive study of SVs in a large series of 762 108 multiple myeloma patients. We show that recurrent and rare SVs are critical in shaping 109 the genomic landscape of multiple myeloma, including complex events simultaneously 110 causing multiple drivers. 111 6 112
Results
113
Genome-wide landscape of structural variation in multiple myeloma 114To define the landscape of simple and complex SVs in multiple myeloma, we 115 investigated 762 newly di...