Supplementary Information

Transcription

1 Supplementary Information Materials/Subjects and Methods Patients The study was approved by the local ethics committee (REC No: 09/H0606/5). All patients gave written informed consent in accordance with the Declaration of Helsinki. Patients followed up at Liverpool University Hospital, Bournemouth Hospital and the Oxford Radcliffe Cancer and Haematology Centre were recruited into the study. Patient characteristics are summarized in Table 1. Overall, the patient cohort was that of a tertiary referral centre with a high proportion of unmutated and treatment cases. Sample collection To minimize ambiguities in OncoSNP 1 data interpretation as a result of sample contamination, only samples from patients with lymphocytes contributing >90% of the total white blood cell count (WBC) were collected. To achieve this, samples were analyzed by multiparameter flow cytometry and expert hematopathological review. Samples with <90% circulating B-CLL cells were excluded. These represented <10% of patients, including patients with small lymphocytic lymphoma and early stage patients. In total, 93 pre-treatment samples were collected.16 patients did not require any treatment, 18 had chemotherapy disease and for 15 patients, follow-up was <1 year. 48 patients relapsed and for 42 of these, follow-up samples before treatment of subsequent relapse were also available. Time to relapse varied from patient to patient (6 months to 5 years), but follow-up samples were always from the next relapse. DNA was extracted from vital frozen cells using the QIAamp DNA Midi Kit (Qiagen, Hilden, Germany) according to the manufacturer s protocol. A flowchart documenting the overall study design, the numbers of samples involved and where they fit into the work undertaken is given in Supplementary Fig 1. Clinical Risk Score and Statistical Analyses To investigate the clinical significance of CNAs/cnLOHs, we assigned a clinical risk score to each clinically evaluable patient in our cohort (n=76). Scores assigned were were: 0=on watch and wait (W&W), 1=progression free survival (PFS) >1year, 2=PFS<1year and 3=chemotherapy ). We also categorized additional covariates for testing (see Summary Table for Statistical Testing below). We used 1

2 the total length of CNAs noted rather than the area under the curve (AUC) because B-CLL is relatively stable and CNAs are comparatively sparse, confined mainly to single copy number deletions or gains. Thus the AUC for B-CLL data reduces down to a measure of total aberration length. This is in contrast to highly unstable cancers such as polyploidy breast cancer or chromosomally unstable colon cancer, where high values of AUC will result from long abnormalities with high copy number. Logistic regression was used to evaluate the association between one response and several prognostic categories using the appropriate statistical test for ordinal variables eg Kruskal-Wallis test 2, Cochran-Armitage trend exact test 3,4, Chi-square test 5. The significance level for all statistical tests was The analyses were performed using the statistical software SAS 9.2. Summary Table for Statistical Testing Covariate Description Type Categories Patient No ID - - Age Age in years Continuous - Sex Sex Categorical Male, Female CRS Clinical Risk Score Categorical 0, 1, 2, 3 (W&W, PFS>1year, PFS<1year, chemotherapy ) Length of CNAs/cnLOHs Total Length of CNAs/cnLOHs Categorical <1Mb, 1-5Mb, >5 Mb Number of CNAs/cnLOHs Total Number of CNAs/cnLOHs Categorical 0, 1, 2, 3, >3 del11q22.3 Known Abnormality Categorical Yes, no Trisomy12 Known Abnormality Categorical Yes, no del13q14.3 Known Abnormality Categorical Yes, no del17p13.1 Known Abnormality Categorical Yes, no TP53 gene Mutation status Categorical Negative, Positive del17p13.1 or TP53 Known Abnormality Categorical Yes, No IGHV gene status Variable Gene Sequence Categorical Unmutated, Mutated RB1 gene RB1 deletion or cnloh Categorical Yes, No Clonal evolution Clonal evolution Categorical Yes, No CRS = Clinical Risk Score CNA = Copy Number Alteration IGHV = immunoglobulin heavy chain variable gene PFS = Progression Free Survival Fluorescence in situ hybridization (FISH) Interphase cells from CLL cultures were analyzed by FISH using the Vysis CLL FISH panel probe set (Abbott Molecular, Illinois, USA) according to manufacturer s instructions. The slides were viewed using an Axioscope fluorescence microscope (Zeiss, New York, USA) equipped with the appropriate filters and imaging software. A total of 200 nuclei were scored for each probe. Array Studies using the Agilent Technologies 244K genome-wide oligonucleotide platform These hybridizations were carried out according to Agilent protocols found on the Agilent Technologies website 2

3 ( e=productliterature&pageid=1465). Illumina Genome-wide SNP Platform hybridizations Hybridization to Illumina 1M-Duo SNP chips was performed according to manufacturer s protocols found on registration at Due to product discontinuation, 22 samples were hybridized to alternative HumanOmni1-Quad SNP chips according to manufacturer s protocols found on registration at quad_beadchip_kit/documentation.ilmn. In brief, patient DNA was denatured, amplified and enzymatically fragmented and then hybridized onto the BeadChips by rocking in an Illumina Hybridization Oven at 48 C for 16-24hrs. The BeadChips were washed according to the Illumina Inc protocol and the hybridized DNA detected by primer extension with labeled nucleotides followed by detection using fluorescent antibodies. The data were processed using GenomeStudioV (Illumina, Inc., San Diego, California, USA) and then analyzed using OncoSNP v1.0 (see next section). For visual comparisons, the data were processed also using Nexus 5 Discovery Edition (BioDiscovery, Inc., El Segundo, California, USA) with the following settings in the first instance (SNPRank Segmentation): Significance Threshold 1x 10-5; Max Contiguous Probe Spacing (Kbp) ; Min number of probes per segment 5; High Gain 0.6; Gain 0.2; Loss -0.2; Big Loss -1.0; 3:1 sex chromosome gain 1.2; Homozygous Frequency threshold 0.95; Homozygous Value Threshold 0.8; Heterozygous Imbalance Threshold 0.4; Minimum LOH Length (Kb) 20; percentage outliers to remove 3%. All Nexus processed plots were also inspected visually to scan for changes not identified using these analysis settings. Statistical Computational Tool for Data Analysis A number of dedicated cancer tools have emerged recently for SNP array analysis However, not all approaches are best suited for our study. For example, Navin et al. (2010) 9 and Letouze et al. (2010) 10 developed explicit modeling strategies and algorithms to model the sequence of events leading to the genomic aberrations observed in tumor samples. These methods rely on obtaining multiple spatiotemporal samples (e.g. pre-treatment biopsies, sections from different parts of a tumour, metastases, etc) from an individual patient and using the information to infer possible models of tumour progression. In our study, these approaches would be less useful since we have data generated from two time points only and spatial 3

4 heterogeneity is not applicable in this scenario. Furthermore, the degree of genomic complexity in B-CLL is substantially lower than in the breast and bladder cancers considered in other studies. Importantly, we were interested not only in the accumulation of genomic aberrations over time (which in B-CLL is much rarer than the other cancers), but we were also interested particularly in the expansion of distinct sub-clones. For these reasons we selected OncoSNP to detect genomic aberrations and quantify cellular admixture at a per-snp level. We did not develop formal algorithms for elucidating the actual sub-clonal structure; this will be addressed in future work that will be aided by high-depth next generation sequencing technology. OncoSNP Analysis Detailed methods have been described previously 1 (see also OncoSNP was run for 15 EM iterations in intra-tumor heterogeneity mode using a thinning setting of 30 for training. Before analysis, we removed all SNP and monomorphic probes in WTCCC2 germ-line copy number variant regions. In addition, we also removed probes showing high levels of signal variation (greater than three times the average median absolute deviation across all probes) in the WTCCC2 samples. We used natural logarithms when calculating the log Bayes Factor. Putative genomic aberrations detected by OncoSNP with log Bayes Factor less than 30 were removed from consideration and considered to be insignificant. The log Bayes Factor threshold value was required due to unexplained, spatially correlated error artifacts in the data that resemble and occur on the scale of real genomic aberrations (particularly when we allow tumor heterogeneity and small signal deflections that may correspond to aberrations present at low quantities in the sample). As described previously for QuantiSNP 11, the threshold value of 30 allows us to achieve low false positive rates whilst retaining good detection power based on simulation and real data testing. Reference data To help identify and exclude germline CNVs of unlikely relevance from our data we excluded SNP and monomorphic copy number probes in known germline CNV regions from the OncoSNP analysis and we also used the Database of Genomic Variants (DGV), a comprehensive catalog of structural variation noted in control data from many studies and representing data from over 3000 samples 12. This allowed us to define somatic gains and losses >20Kb in size. Copy number changes that encompassed fully changes noted in the DGV were excluded from further analysis. 4

5 However, the DGV does not inform upon cnloh. To determine whether regions of cnloh >2Mb noted in our B-CLL samples were observed commonly in controls, we used data from 150 controls, selected randomly from 6000 control sample data sets. The latter were generated from 3000 samples from the 1958 British Birth Cohort and 3000 from the UK Blood Service Control Group by the Wellcome Trust Case Control Consortium using an Illumina 1.2M custom chip. Although the occurrence of a region of cnloh both in WTCCC2 controls and in B-CLL samples could not rule out the existence of relevant gene mutations in these regions, the information did provide an element of stringency for cnloh selection/exclusion. The main criterion for excluding a cnloh region >2Mb from our data was if it occurred/overlapped relatively frequently in the control samples. By contrast, if a region of cnloh was not observed at all, then we included it in our data because we considered it unlikely to occur at appreciable frequency in the rest of the WTCCC2 cohort. Therefore, for our purposes, 150 data sets (one and a half times the number of pre-treatment samples) were considered sufficient. In addition, we considered CNA analysis against patient constitutional DNA. However, this was not feasible for this study as germline samples were not available; routine collection has started only recently. Furthermore, additional tests would be limited severely by financial constraints Array Data Analysis The Illumina raw data generated using the B-CLL samples are available on request from the Corresponding Author. All CNAs 20Kb and all cnlohs 2Mb were included in the analysis; both were easily verified visually affording a high level of confidence in the results. The comparatively higher 2Mb size threshold for cnloh was applied because very small regions of cnloh are extremely common and current lack of knowledge would render further investigation of putative clinical relevance unviable. Smaller CNAs/cnLOHs were only included if they were in specific regions of interest and visually verified. Imbalances that had been repeatedly observed previously either in control data deposited in the DGV or in the WTCCC2 samples were not pursued further. Similarly, cnloh regions identified repeatedly (identical or overlapping) in the WTCCC2 samples were not pursued further. Where possible, the start and end point positions of CNA/cnLOH regions were defined using OncoSNP. Of interest, we observed that OncoSNP and Nexus calls were highly comparable; the start and end base positions differed subtly, but were 5

6 too small to alter the genes mapping within the minimally deleted regions (MDRs) and minimally overlapping regions (MORs). Where there were apparent discrepancies, the events and positions were manually verified using Nexus. In order to delineate MDRs/MORs, OncoSNP (or in some instances, Nexus) start and end positions of all CNA/cnLOH regions were entered first into Microsoft Excel 2007 spreadsheets. Second, CNAs noted in control samples in the DGV or cnloh regions occurring commonly in the WTCCC2 data were removed. Third, regions of overlap involving the remaining CNAs/cnLOHs were identified using Microsoft Excel 2007 functions and the start and end positions of MDRs/MORs identified according to the start/end positions of the changes noted in informative samples. To be classified as recurrent, CNA/cnLOH regions had to occur at least two times, either in different pre-treatment samples or once in a pre-treatment sample and once newly emerging in a different relapse sample. Where relapse samples showed an increased percentage of cells compared with the pre-treatment samples, this was also noted. To increase the stringency for identifying MORs harboring putative driver genes, MORs >5Mb were excluded. To avoid any inclusion of germ-line variants, CNAs kb were only included if they showed mosaicism, if they were homozygous deletions involving at least one exon, or if they mapped to a known region of interest or had newly emerged/showed increase in percentage at relapse. Results are given in Table 2. Pathway Analysis We identified 706 genes within the 62 MORs presented in Table 2. After exclusion of genes mapping to regions with known copy number variants reported in the DGV, 542 genes remained and were tested for pathway enrichment using R/Bioconductor org.hs.eg.db and GOstats packages against a universe of 45,469 possible IDs. We noted KEGG Pathways significantly over-represented at p Results were validated using the DAVID web-based interface 13. These analyses were then repeated adding RB1, ATM, TP53 and DLEU2 to the previous list of genes. TP53 mutation analysis TP53 was screened for mutations using four high-resolution melting (HRM) assays covering exons 5-8. Primers sequences are provided in the Table below. Each 10ul PCR reaction contained 1x HotShot Diamond (Clent Life Sciences, Stourbridge, UK) 1x LCGreen (Clent Life Sciences), 1x Q solution (Qiagen), 1uM each primers and 40ng DNA. The PCR amplification consisted of an initial enzyme activation step of 95 C for 10 min, followed by 45 cycles of 95 C for 15sec, 60 C for 20sec, 72 C for 6

7 15sec. This was followed by the HRM which increased from 72 C to 95 C by 0.1 C every 2sec. Both amplification and HRM were performed on a Rotorgene 6000 (Qiagen). Following HRM results were analyzed on the HRM module of the Rotor- Gene 6000 software 1.7. Positive PCR products were purified then sequenced from both strands. The products were analyzed on the CEQ 8000 (Beckman Coulter, California, USA) using the Sequence Investigator software. Primers used for TP53 sequencing Primer Sequence (5'-3') TP53 exon 5 (F) TGCCGTCTTCCAGTTGCT TP53 exon 5 (R) CATCGCTATCTGAGCAGC TP53 exon 6 (F) CAGGCCTCTGATTCCTCACTGATTGCTC TP53 exon 6 (R) CTCACCTGGAGGGCCACTGACAA TP53 exon 7 (F) CAAGGCGCACTGGCCTCATCT TP53 exon 7 (R) CAGGCCAGTGTGCAGGGT TP53 exon 8 (F) GGTAGGACCTGATTTCCTTACT TP53 exon 8 (R) AGGCATAACTGCACCCTTGG Analysis of Immunoglobulin heavy chain variable (IGHV) gene rearrangements To identify clonal rearrangements of the immunoglobulin heavy (IGH) chain gene and determine the somatic mutation status of the variable (V) gene sequence in patients we used the IGH Somatic Hypermutation (SHM) Assay v2.0 (Invivoscribe, La Coutat, France) according to the manufacturer s instructions. Sequences were run on the CEQ 8000 (Beckman Coulter) and analyzed by Vquest ( The IGHV gene status of a patient was categorized as mutated if there was <98% homology with germline and unmutated if there was 98% homology with germline or presence of the V3-21 rearrangement 14. BLIMP1 sequencing 11 assays were designed to cover the whole of the coding region of the BLIMP1 gene. Primer sequences are provided in the Table below. Each primer had a M13 tail; forward: 5 GTAAAACGACGGCCAGT 3 and reverse: 5 CAGGAAACAGCTATGAC 3. Each 25µl reaction contained 1x Qiagen MasterMix (Qiagen) and 10mM of each primer. All PCR conditions, except for 5b, consisted of an initial enzyme activation step of 95 C for 15 min, followed by 35 cycles of 95 C for 30sec, 60 C for 90sec, 72 C for 90sec, followed by a final 10min extension at 72 C. Conditions for 5b were the same except the annealing temperature was 57 C. Purified amplicons were sequenced from both strands using the M13 tails, then 7

8 analyzed on the CEQ 8000 (Beckman Coulter, California, USA) using the Sequence Investigator software and compared with the corresponding germ-line sequences. Primers used for BLIMP1 sequencing Primer Sequence (5'-3') BLIMP1 exon 1 (F) GCCGAGTGGCTAAGGAAATC BLIMP1 exon 1 (R) GCAGAAGCATGTTTCTACTGC BLIMP1 exon 2 (F) CATAGCCTCTCAGAAGGAGC BLIMP1 exon 2 (R) GTAGGGAGATTTGGCACCAG BLIMP1 exon 3 (F) GGCATCATTAATGTCTGTTTACTTATC BLIMP1 exon 3 (R) CGCACCTTAGTCCCAGCTAC BLIMP1 exon 4 (F) CTTACCTGTTTCCGCCCTG BLIMP1 exon 4 (R) CAGCATGTCTGGACCAATCC BLIMP1 exon 5a (F) CTAGCCCTCTGTGTAATCGC BLIMP1 exon 5a (R) GTTGTTGATGCCATTCATGC BLIMP1 exon 5b (F) AGTACGCTCACTACCCCAAGTTC BLIMP1 exon 5b (R) CTCACAGCCCCTTGGACTG BLIMP1 exon 6 (F) GAGCCAGCTTGAGAGCAGAG BLIMP1 exon 6 (R) TGGGAGGGTGACTCACAGAC BLIMP1 exon 7a (F) CCCGTTGGCAACTCTTAATC BLIMP1 exon 7a (R) GCCTTTGCCTTGTTCATGC BLIMP1 exon 7b (F) AAATGGTTTCCCCTCACCTC BLIMP1 exon 7b (R) TGGGGAAATTTTCGCAGTG BLIMP1 exon 7c (F) AGTCCTGTGGCCATTCAGAG BLIMP1 exon 7c (R) ACCGACGTGATTGTGAGGTC BLIMP1 exon 7d (F) TGAGCCAAGCCATGTAAAAG BLIMP1 exon 7d (R) TGGAGATGTCGCTCACTGAC ATG5 sequencing 10 assays were designed to cover the whole of the coding region of the ATG5 gene. Primer sequences are provided in the Table below. Each primer had a M13 tail; forward: 5 GTAAAACGACGGCCAGT 3 and reverse: 5 CAGGAAACAGCTATGAC 3. Each 25µl reaction contained 1x Qiagen MasterMix (Qiagen) and 10mM of each primer. All PCR conditions consisted of an initial enzyme activation step of 97 C for 15 min, followed by 35 cycles of 97 C for 30sec, 60 C for 90sec, 72 C for 2 min, followed by a final 10min extension at 72 C. Purified amplicons were sequenced from both strands using the M13 tails, then analyzed on the CEQ 8000 (Beckman Coulter, California, USA) using the Sequence Investigator software and compared with the corresponding germ-line sequences. 8

9 Primers used for ATG5 sequencing Primer Sequence (5'-3') ATG5 exon 1 (F) CTGGGTTAGGCAGAACACG ATG5 exon 1 (R) ACAAGGTGGACACACACACG ATG5 exon 2 (F) TGCAAGGATCTGACTAATGCTC ATG5 exon 2 (R) CCCATTTGCCACAATCAATG ATG5 exon 3 (F) AAGAACACGGCTGTTTTTCC ATG5 exon 3 (R) TGCTTAATAATGCAGAAAAATTCAC ATG5 exon 4 (F) TTAAAGCCCCTGACATTTGG ATG5 exon 4 (R) AATGGGACGAAGGAGAAATG ATG5 exon 5 (F) TTGAAAAACTGGGGGATATAGTTC ATG5 exon 5 (R) AATCTGGGCACAGAGGCTAC ATG5 exon 6 (F) TTTCATCTCTTCATGTGAGGTATTC ATG5 exon 6 (R) TGTCTGAGGCTTTCATAAATGG ATG5 exon 7 (F) AAAAGGCACCTAATGCCAAC ATG5 exon 7 (R) GAAATGTTTTAATGTTGCTGATTG ATG5 exon 8a (F) TTGTTGGGTTTCTTTCTTGG ATG5 exon 8a (R) TTCGTTAAGGAAAGATGGGTTTAC ATG5 exon 8b (F) TTGAACTTTAGCTCATGAAAGTGG ATG5 exon 8b (R) TCAGTGAAAATCGCAAAAGG ATG5 exon 8c (F) CGATCATGGTTTTAGATCCCATA ATG5 exon 8c (R) TTTTAAATAAAGACGGACACAACA Results Comparison between FISH and Agilent Technologies 244K and Illumina 1M-Duo array platforms As part of the validation of array based platforms for the detection of CNAs in CLL, we tested 31 patient DNAs using both the Agilent 244K and the Illumina 1M-Duo platforms and compared the results with those obtained by routine FISH (Supplementary Table 1). Both array platforms detected all of the CNAs seen by FISH. Importantly, one case presented in Supplementary Table 1 (CLL034) and an additional case tested only on the Illumina Platform (CLL118) had small 11q deletions encompassing the ATM gene that were not noted in the patients FISH results. The array data show that the ~732kb region interrogated by the probe from the Vysis ATM/CEP11 FISH Probe Kit overlaps proximally with a non-deleted region in both cases, thereby allowing probe binding (Supplementary Fig. 2). Both array platforms also detected five 13q deletions that were not reported in the FISH results (Supplementary Table 1). In addition, the use of arrays enabled the extent and boundaries of the abnormalities to be defined more accurately. 9

10 Choice of Platform Whilst both Agilent and Illumina platforms allowed the reliable detection of CNAs, only the SNP-based Illumina platform was capable of identifying regions of cnloh that might be important in B-CLL evaluation. For example, three out of 93 patients, CLL071, CLL075 and MCLL148, had cnloh encompassing TP53. Mutation analysis of these patients revealed TP53 mutations on both alleles (Supplementary Table 2, Supplementary Fig. 3). When considering which SNP platform to choose for the study, the Illumina platform was preferred to the Affymetrix because of the superior signal to noise ratio. The content of the 1M-Duo chip (the highest density chip available for use at the time) matched, in large part, the 1M-Duo design used for the WTCCC2 study (1,109,332 probes in common). Therefore, this platform was chosen to extend our studies. Validation of OncoSNP Quantification Spiking experiments have shown previously the excellent precision of the quantitative data obtained with OncoSNP using promyelocytic leukemia cancer, colon cancer and breast cancer samples 1. To further validate the quantifications obtained, we compared the OncoSNP output for heterozygous losses with visual inspection of Nexus plots; the percentage calls were consistent with the observations (Supplementary Fig. 4). Furthermore, we performed blindly a series of interphase FISH experiments using cells from a subset of patients with known cytogenetic abnormalities involving 11q22.3, 13q14.3, or 17p13.1 (Supplementary Table 3). This comparison showed good correlation in 18/20 samples tested. One of the two potentially inconsistent cases was CLL071, which involved a homozygous del13q14.3. Here the OncoSNP call of 100% was not stated in Supplementary Table 3 because formal OncoSNP calibration for homozygous losses is pending. Thus, inaccurate calibration may be the reason for the inconsistency in this case. The second inconsistent result identified was for CLL084 where OncoSNP did not detect the 15% 13q14.3 loss reported by FISH. On visual inspection of the array data, we were unable to confirm the CNA despite the fact that a low-level CNA of 15% would be expected to be visible. As FISH background levels of 10%-20% are allowable, we believe that the most likely explanation here is that the FISH result may reflect nonspecific background. 10

11 Limitations of OncoSNP in this Study Ultimately, the accuracy with which we can quantify cellular admixture using OncoSNP or any other computational tool is determined by the noise variances of the SNP data. Whilst OncoSNP was essential for our study, the performance is subject to technological limitations which mean it is not proven accurate enough to report mixed cell populations of >90% or <10% and has not been calibrated formally for homozygous deletions. However, for some patient samples we did record values >90% (but<100%), or of apparently 100% based on visual inspection of Nexus B- allele frequency plots (Supplementary Fig. 4). For six cases, an aberration change from 90% to >90% or vice versa was noted between paired pre-treatment and relapse samples, potentially impacting on interpretation of clonal evolution. For these cases, the possibility of variation in sampling (namely variation in the percentage of B-cells harvested from paired pre-treatment sample and relapse samples) had to be taken into consideration as a possible explanation for the observations. For three of the six cases noted it was not possible to rule out sampling variation on the basis of the array data alone. However, for the remaining three cases, comparison with CNAs from other chromosomal regions of the samples suggested strongly that the differences in percentages were not attributable to sampling. For example, in both the pre-treatment and relapse samples for CLL068, the percentage of cells carrying the chromosome 13q loss remained unchanged, whilst the percentage of cells carrying the chromosome 2 loss increased visibly from 90% in the pre-treatment sample to >90% in the relapse sample (data not shown). Finally, we observed that some cnloh regions visualized manually and reported here, were interpreted by OncoSNP as germ-line variants. We expect that the fundamental limits of analogue array based technology will be addressed ultimately by digital high resolution next generation sequencing technologies. However, for this study, our experience was that the data were best interpreted by using the output of OncoSNP in combination with manual visualization and verification of results using software such as Nexus. Analysis of well-recognized CNAs in 93 pre-treatment samples 11q22.3 and 17p13.1 abnormalities We identified a total of 17 out of 93 patients with deletions 500kb involving 11q22.3 (one of these, in CLL084, is newly visible at relapse) and two with cnloh >2Mb). As outlined above, for two cases, CLL034 and CLL118, the deletions were missed by 11

12 FISH using the standard Abbott/Vysis probes (Supplementary Fig. 2). Including these cases, the MDR in our sample set was ~577kb involving CUL5, ACAT1, NPAT, ATM, C11orf65, KDELC2, EXPH5 and DDX10. In both patient and control groups there were small regions of cnloh that included ATM, but two patients showed large regions of cnloh that were not observed in controls; CLL090 gave a mosaic pattern showing 50% cnloh 11qcen-qter, whereas CLL108 showed complete cnloh for 11q12.3-qter. Although we did not formally prove the presence of ATM mutations in these patients, the results point to a possible role of cnloh in CLL. Nine of the 93 samples tested carried single copy losses involving 17p13.1. Eight of these had TP53 mutations of the non-deleted allele; three of the 93 had mono-allelic mutations. A further three patients had bi-allelic mutations with cnloh encompassing TP53 (Supplementary Table 2, Supplementary Fig. 3). None of the control samples showed CNAs or cnloh of this region. This data provides further evidence that areas of cnloh can indicate hotspots for pathogenically important mutations. A further 2/93 patients (CLL093 and CLL108) had mono-allelic mutations. When we tested statistically the status of TP53 mutations vs CRSs we found that patients who were TP53 mutation positive showed significantly higher CRSs (80.0% 2 CRS) than TP53 mutation negative patients (46.6% 2 CRS) (Cochran-Armitage trend exact test, p-value=0.0154; Supplementary Table 5). Statistically significant differences were also found with respect to the total length of CNAs/cnLOHs between TP53 mutation groups (Cochran-Armitage trend exact test, p-value=0.0081; Supplementary Table 5). 13q14.3 abnormalities The most common abnormalities were CNAs/cnLOHs involving 13q14.3, detected in 47/93 of the pre-treatment patients. The ~7kb MDR in these cases involved DLEU2. However, no single patient had a 13q abnormality restricted to DLEU2 (Supplementary Fig. 5). 33 patients had heterozygous deletions encompassing the MDR whereas 14 had a homozygous deletion involving this region (Supplementary Table 6). In 24 patients, the deletion included the retinoblastoma gene RB1 at 13q14.2. Seven patients with 13q abnormalities showed large regions of cnloh. One of them (CLL074) showed isolated cnloh involving the 13q14.3 MDR. Quantification of the subclonal distribution of CNAs/cnLOHs involving the 13q locus revealed that whilst the MDR was deleted in the majority of leukemic cells in individual patients, regions next to the MDR were deleted only in subpopulations of the same leukemia (eg: CLL060). 12

13 Together, these data confirm previous observations that the 13q14.3 locus is highly heterogeneous in CLL patients. Furthermore, the results suggest that apparently smaller abnormalities involving the MDR can gradually extend through clonal evolution from an initially small proportion of cells carrying the extended abnormality (Supplementary Fig. 6). 13q abnormalities and clinical outcome There has been accumulating evidence that the size of 13q deletions has prognostic relevance with regards to depth and duration of response to treatment However, recent genetic evidence from conditional knock-out mice suggests that deletion of the MDR containing the mir15a/16-1 locus is sufficient to control the expansion of the mature B-cell pool in mice leading to a CLL-like disease albeit with low penetrance and after a long period of latency 18. In an attempt to establish whether there was an association between the size of the 13q deletion and clinical outcome, we compared the size of 13q deletions with the clinical risk score. To this end, we defined Type I and Type II del13q14.3 as previously described on the basis of aberrations affecting the RB1 gene (16). We excluded patients with additional known CNAs/cnLOHs in addition to 13q deletions (in order to eliminate compounding factors) and patients for whom no CRS was available and analyzed only patients who had del13q as their sole abnormality (21/47 patients). When we looked at CRSs in patients without RB1 deletions or cnloh compared with those with RB1 deletions or cnloh we found that patients without RB1 deletions or cnlohs showed higher CRSs (73.7% 2 CRS) than patients who carried RB1 deletions or cnlohs (42.1% 2 CRS) (Cochran- Armitage trend exact test, p-value=0.0466; Supplementary Table 7). Thus, in our small cohort, large Type II 13q deletions were not associated with a worse clinical outcome. When we looked at IGHV gene status (unmutated/mutated) in patient samples with isolated 13q anomalies we noted that patients without RB1 deletions or cnlohs showed a slightly higher percentage of cases with mutated IGHV genes (60.0%) compared with patients with RB1 deletions or cnlohs (41.2%). However, this did not reach statistical significance (Chi-square test, p-value=0.2879; Supplementary Table 7). Finally, we also found recurrent deletions of 22q involving the PRAME locus 19. However, both large and small deletions involving this locus were also observed in the WTCCC samples indicating that this deletion is a common variant. Therefore it was excluded from further analysis. 13

14 Patients without apparent genomic complexity Recent studies of acute myeloid leukemia have shown that a large proportion of cases exhibit low genomic instability, with 50% cases showing no acquired CNAs or cnloh at the resolution tested 20. By contrast, when we investigated infrequent CNAs/cnLOH events in our B-CLL pre-treatment cases (using 20kb and 2Mb thresholds respectively), we found that only ~4% cases had no CNAs/cnLOH events identified, ~18% cases had one CNA (including three cases of trisomy 12) and ~17% had two CNAs (including three different cases with trisomy 12). Pathway Analysis When we tested the 542 genes from the MDRs/MORs of interest using the KEGG pathway analysis, we obtained ten pathways over-represented at a P-value<0.05. When ATM, RB1, DLEU2 and TP53 were added to the gene list we obtained 17 pathways over-represented at a P-value<0.05, many of these with a strong link to cancer (Supplementary Table 10). Other pathways of interest were also identified, though did not reach statistical significance in this analysis. The results for the DAVID analyses were similar in each case. 14

15 Supplementary References 1. Yau C, Mouradov D, Jorissen RN, Colella S, Mirza G, Steers G, et al. A statistical approach for detecting genomic aberrations in heterogeneous tumor samples from single nucleotide polymorphism genotyping data. Genome Biol 2010 Sep 21; 11(9): R Kruskal WH and Allen Wallis, W. Use of ranks in one-criterion variance analysis. Journal of the American Statistical Association 1952 Dec; 47(260): Cochran, WG Some methods for strengthening the common chi-square tests. Biometrics 1954 Dec; 10 (4): Armitage, P. Tests for Linear Trends in Proportions and Frequencies. Biometrics 1955 Sep; 11 (3): Pearson K. On the criterion that a given system of deviations from the probable in the case of a correlated system of variables is such that it can be reasonably supposed to have arisen from random sampling. Philosophical Magazine, Series ; 50 (302): Greenman CD, Bignell G, Butler A, Edkins S, Hinton J, Beare D, et al. PICNIC: an algorithm to predict absolute allelic copy number variation with microarray cancer data. Biostatistics 2010 Jan; 11(1): Sun W, Wright FA, Tang Z, Nordgard SH, Van Loo P, Yu T, et al. Integrated study of copy number states and genotype calls using high-density SNP arrays. Nucleic Acids Res 2009 Sep; 37(16): Van Loo P, Nordgard SH, Lingjaerde OC, Russnes HG, Rye IH, Sun W, et al. Allele-specific copy number analysis of tumors. Proc Natl Acad Sci U S A 2010 Sep 28; 107(39): Navin N, Krasnitz A, Rodgers L, Cook K, Meth J, Kendall J, et al. Inferring tumor progression from genomic heterogeneity. Genome Res 2010 Jan; 20(1): Letouze E, Allory Y, Bollet MA, Radvanyi F, Guyon F. Analysis of the copy number profiles of several tumor samples from the same patient reveals the successive steps in tumorigenesis. Genome Biol 2010; 11(7): R Colella S, Yau C, Taylor JM, Mirza G, Butler H, Clouston P, et al. QuantiSNP: an Objective Bayes Hidden-Markov Model to detect and accurately map copy number variation using SNP genotyping data. Nucleic Acids Res 2007; 35(6): Iafrate AJ, Feuk L, Rivera MN, Listewnik ML, Donahoe PK, Qi Y, et al. Detection of large-scale variation in the human genome. Nat Genet 2004 Sep; 36(9):

16 13. Wei Huang D, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature Protocols 2009; 4(1):44 & Genome Biology 2003; 4(5):P3 ). 14. Oscier DG, Gardiner AC, Mould SJ, Glide S, Davis ZA, Ibbotson RE, et al. Multivariate analysis of prognostic factors in CLL: clinical stage, IGVH gene mutational status, and loss or mutation of the p53 gene are independent prognostic factors. Blood 2002 Aug 15; 100(4): Ouillette P, Erba H, Kujawski L, Kaminski M, Shedden K, Malek SN. Integrated genomic profiling of chronic lymphocytic leukemia identifies subtypes of deletion 13q14. Cancer Res 2008 Feb 15; 68(4): Parker H, Rose-Zerilli MJ, Parker A, Chaplin T, Wade R, Gardiner A, et al. 13q deletion anatomy and disease progression in patients with chronic lymphocytic leukemia. Leukemia 2011 Mar; 25(3): Ouillette P, Collins R, Shakhan S, Li J, Li C, Shedden K, et al. The prognostic significance of various 13q14 deletions in chronic lymphocytic leukemia. Clin Cancer Res Sep Klein U, Lia M, Crespo M, Siegel R, Shen Q, Mo T, et al. The DLEU2/miR- 15a/16-1 cluster controls B cell proliferation and its deletion leads to chronic lymphocytic leukemia. Cancer Cell Jan 19; 17(1): Gunn SR, Bolla AR, Barron LL, Gorre ME, Mohammed MS, Bahler DW, et al. Array CGH analysis of chronic lymphocytic leukemia reveals frequent cryptic monoallelic and biallelic deletions of chromosome 22q11 that include the PRAME gene. Leuk Res 2009 Sep; 33(9): Walter MJ, Payton JE, Ries RE, Shannon WD, Deshmukh H et al. Acquired copy number alterations in adult acute myeloid leukemia genomes. Proc Natl Acad Sci Aug 4; 106(31):

17 Supplementary Table 1. Comparison between FISH and Agilent Technologies and Illumina Inc. array results* Patient No Reported FISH result Agilent 244K Illumina 1M-Duo CLL010 normal normal normal CLL026 normal normal normal CLL034 del13q14.3 del11q22.3; del13q14.2-q14.3 del11q22.3; del13q14.2-q14.3 CLL035 trisomy 12 trisomy 12 trisomy 12 CLL038 del11q22.3 del11q del11q CLL041 del13q14.3 homo del13q14.12-q14.3 (homo del13q14.2-q14.3) del13q14.12-q14.3 (homo del13q14.2-q14.3) CLL042 trisomy 12 trisomy 12 trisomy 12 CLL043 del17p13.1 del17pcen-pter 90% del17pcen-pter CLL044 del17p13.1 del17pcen-pter; homo del13q14.3 del17pcen-pter; homo del13q14.3 CLL046 normal del13q14.2-q14.3 del13q14.2-q14.3 CLL047 normal normal normal CLL048 trisomy 12 trisomy 12 trisomy 12 CLL050 del13q14.3 del13q14.2-q14.3 del13q14.2-q14.3 CLL055 del11q22.3 del11q14.1-q % del11q14.1-q23.2 CLL056 del13q14.3 del13q % del13q14.3 CLL058 del17p13.1 del17p13.1; del13q14.2-q14.2 with homo del13q14.2-q % del17p11.2-pter; del13q14.2-q14.2 with homo del13q14.2-q14.3 CLL060 del13q14.3 del13q % del13q14.3 CLL063 del11q22.3 homo del13q14.3; del11q22.3 cnloh chr13q12.11-qter wth homo del13q14.3; del11q CLL064 normal del13q14.3 (mosaic) 50% del13q14.3 CLL071 homo del13q14.3 homo del13q14.3 (MDR) 70% del13q14.2-q14.3 with homo del13q14.3 (MDR); 30-70% cnloh 17p11.2-pter CLL076 trisomy 12 trisomy 12 80% trisomy 12 CLL077 normal normal normal CLL079 normal normal normal CLL080 del13q14.3 del13q14.3 (mosaic) with del13q14.3 MDR ~50% del13q14.3 with del13q14.3 MDR CLL081 del17p13.1;del13q14.3 del17pcen-pter; del13q14.3 (mosiac) 90% del17pcen-pter; 50% del13q14.3 CLL082 homo del13q14.3 del13q with homo del13q14.3 del13q with homo del13q14.3 CLL083 normal normal normal CLL084 normal normal normal CLL085 del13q14.3 del13q14.11-q21.2 (mosaic) 50% del13q14.11-q21.2 CLL087 trisomy 12 trisomy 12 80% trisomy 12 CLL088 del13q14.3 del13q14.3 >90% del13q14.3 * Results given are for the 11q22.3, 13q13.4 and 17p13.1 FISH probe regions of pre-treatment samples and 'normal' refers only to these regions. All deletions are heterozygous unless stated as homozygous ('homo'). Illumina percentages represent CNA/cnLOH levels determined using OncoSNP. Where a percentage is not noted for the Illumina result, then all cells appear to carry the anomaly Grey shading highlights samples showing differences between FISH and array analysis. Samples with cnlohs identifiable only using Illumina Platform 17

18 Supplementary Table 2. Clinical and Molecular Characteristics of B-CLL patients included in the study Patient No Age (yrs) Sex (M/F) IGHV gene status TP53 mutation status Current management PFS CRS CNAs and cnloh events * CLL M NK Negative W&W N/A 0 47,XY, arr 70% 2pterp14(0-65,091,911)x3, 11q22.3(107,477, ,054,389)x1, 13q14.2q14.3(48,737,980-50,553,577)x1 CLL F Mutated Negative W&W N/A 0 47,XX, arr 3p24.3(21,592,730-22,039,016)x1, 12(0-132,146,663)x3 CLL F Unmutated Negative W&W N/A 0 46,XX, arr 6p21.32p22.1(25,791,596-32,491,015)x2 hmz, 80% 10q24.32(103,708, ,851,056)x1, 13q14.12q14.3(45,763,393-47,874,856)x1, 13q14.2q14.3(47,874,856-50,427,794)x0, 13q14.3(50,427,794-50,514,234)x1 CLL M Unmutated Negative W&W N/A 0 47,XY, arr 12(0-132,146,663)x3 CLL M Mutated NK W&W N/A 0 None noted CLL M Mutated Negative W&W N/A 0 49,XY, arr 3p14.2(59,416,360-59,688,740)x3, 12(0-132,146,663)x3, 18(0-76,116,152)x3, 19(0-63,507,665)x3, Xp21.1(31,700,590-31,763,529)x0 CLL M Mutated Negative W&W N/A 0 47,XY, arr 5p13.1(40,238,262-40,263,075)x1,12(0-132,146,663)x3 CLL M NK Negative W&W N/A 0 46,XY, arr 13q12.11qter(22,060, ,121,252)x2 hmz, 13q14.3(49,460,089-50,493,338)x0 CLL M Unmutated Negative W&W N/A 0 46,XY, arr 90% 11q14.1q23.2 (78,697, ,120,247)x1 CLL M NK Negative W&W N/A 0 46,XY, arr 1p33(49,687,041-49,772,789),x1, 7p22.1(4,932,547-5,123,570)x3, 11p13(33,083,227-33,228,359)x3, 50% 13q14.12q14.13(44,532,138-45,941,312)x1, <10% 13q14.2(45,941,312-47,522,891)x1, 50% 13q14.2(47,522,891-48,463,998)x1, 80% 13q14.3(48,463,998-50,493,338)x1, <10% 16p13.3(3,196,980-4,598,585)x1 CLL M Mutated Negative W&W N/A 0 46,XY, arr 80% 6q21 (106,208, ,080,443)x1, 17q22q23.1(54,877,659-55,032,870)x3 CLL M Mutated Negative W&W N/A 0 47,XY, arr 17q22q23.1(54,877,659-55,032,870)x3 CLL M Mutated NK W&W N/A 0 46,XY, arr 3q13.32(120,256, ,302,958)x3, 50% 13q14.2q14.3(46,776,954-50,012,183)x1, 40% 13q14.3(50,012,183-52,794,521)x1 CLL M NK Negative W&W N/A 0 46,XY, arr 13q14.2-q14.3(48,660,957-51,331,565)x1 CLL072 NK M Mutated Negative W&W N/A 0 46,XY, arr 13q14.2q14.3 (48,381,735-50,599,548)x1 CLL F NK Negative W&W N/A 0 46,XX, arr 13q14.3(49,358,002-50,499,548)x1, 90% 13q14.3(49,468,192-50,355,810)x0 CLL F Unmutated Negative pre-flu/cy >1year 1 None noted CLL F Unmutated Negative pre-chlorambucil >1year 1 46,XX, arr 4q24(107,162, ,237,623)x3, 40% 10p12.23(17,833,779-18,398,113)x1, 40% 10q24.32q25.2 (103,562, ,637,185)x1, CLL M Unmutated NK pre-flu/cy >1year 1 47,XY, arr 18p11.32(146, ,422)x3 CLL F Unmutated Negative pre-chlorambucil >1year 1 46, XX, arr 6q14.1 (83,165,917-83,265,032)x1, 7p21.1(18,576,853-18,602,061)x3, 11q12.3(62,082,265-62,330,255)x1, 13q14.2q14.3(48,608,987-51,326,824)x1 CLL M Mutated Negative pre-chlorambucil >1year 1 46,XY, arr 90% 2q22.2q24.1(144,240, ,961,129)x1, 13q14.2q14.3(47,595,821-50,072,312)x1 CLL F Mutated NK pre-chlorambucil >1year 1 46,XX, arr 3q25.31(156,553, ,572,680)x1, 30% 13q14.11q14.3(40,674,362-49,342,922)x1, ~50% 13q14.3(49,342,923-50,332,192)x1 CLL083 NK M Unmutated Negative pre-flu/cy >1year 1 46,XX, arr 4p14(39,001,471-39,049,059)x3, 11q25(131,123, ,186,533)x3, <10% 12q14.3(64,809,667-64,853,817)x1, 12q23.1(99,175,388-99,224,488)x1,10% 16p12.2(20,710,305-20,790,857)x3, 17q12(34,817,937-34,948,998)x3 CLL M Unmutated Negative pre-flu/cy >1year 1 46,XY, arr 10q23.1q23.31(85,670,734-89,454,971)x2 hmz, 50% 13q14.11q21.31(43,934,565-61,368,181)x1 CLL089 NK M Unmutated Negative pre-flu/cy >1year 1 46, XY, arr 1q21.1(144,335, ,614,943)x3, 1q25.2(174,495, ,547,429)x3 CLL F Mutated Negative pre-chlorambucil >1year 1 47,XX, arr 90% 7q33q34(134,397, ,656,577)x1, 90% 7q36 (156,694, ,824,049)x1, 50% 11q12.1qter(60,468, ,445,626)x2 hmz, 80% 12(0-132,146,663)x3, 80% 13q14.3(49,296,907-50,580,804)x1 CLL091 NK M Mutated Negative pre-chlorambucil >1year 1 46,XY, arr 14q24.3q31.1(77,490,883-80,987,659)x2 hmz CLL M Mutated HRM Positive pre-chlorambucil >1year 1 46,XY, arr 80% 6p25.3p25.2-(94,609-4,294,807)x3, 6q14.1-qter(83,755, ,753,209)x1, 70% 10q25.1(109,682, ,770,215)x3, 13q14.2-q14.3(47,556,552-50,610,384)x1, 70% 20q13.13(47,292,495-47,500,772)x3 CLL F Mutated Negative pre-flu/cy >1year 1 46,XX arr 40% 13(0-114,121,252)x2 hmz, 13q14.3(49,425,974-49,522,141)x1, 13q14.3(49,522,141-50,380,713)x0 MCLL F Unmutated Negative pre-fcm-minir >1 year 1 46,XX, arr 4q24(107,259, ,283,621)x1, 12q13.11q24.33(47,384, ,289,191)x2 hmz CLL097 NK M Mutated NK NK >1year 1 46, XY, arr >90% 13q14.11q14.3(40,332,062-51,453,289)x1 CLL098 NK M Unmutated Negative NK >1year 1 None noted CLL099 NK F Unmutated Negative NK >1year 1 46,XY, arr 70% 1q21.1(144,099, ,757,265)x3 CLL100 NK M Unmutated Negative NK >1year 1 46,XY, arr 6q23.3(136,941, ,303,525)x1, 11q21q23.3(95,807, ,159,645)x1, 13q14.12q14.3(45,803,275-50,435,049)x1 CLL101 NK M Unmutated Negative NK >1year 1 46,XY, arr 11q14.1qter(77,080, ,445,626)x1, 60-70% 13q14.2q21.32(47,781,909-64,410,990)x1, 50% 13q21.32q33.2(64,350, ,707,225)x3, 70% 22q11.21qter(17,403,754-49,582,267)x3 CLL105 NK M Unmutated Negative NK >1year 1 46,XY, arr 80% 10q24.31q33(102,616, ,675,986)x1, 90% 11q14.1q24.2(78,235, ,814,884)x1, 90% 13q14.2q14.3(47,373,566-51,972,428)x1, 13q14.3(49,425,974-49,953,648)x0 CLL107 NK M Unmutated Negative NK >1year 1 46,XY, arr 50% 3p21.31(46,827,053-50,133,778)x1, 50% 3p21.1(52,038,099-52,553,933)x1, >90% 11q14.3q24.3(91,864, ,805,084)x1, >90% 13q14.11(40,389,170-40,761,341)x1, >90% 13q14.11(41,687,912-41,826,016)x1, >90% 13q14.2q14.3(47,515,748-51,333,756)x1 CLL108 NK F Unmutated Phe212- SerfsX3 NK >1year 1 46,XX, arr 3q22.3(137,678, ,832,637)x1, 80% 6q14.1q22.33(81,463, ,051,286)x1, 90% 11q12.3qter(62,398, ,445,626)x2 hmz, 80% 12pterp12.2(0-20,356,457)x2 hmz, 80% 12p12.2q21.31(20,356,457-84,251,286)x3, 80% 12q21.31qter(84,251, ,146,663)x2 hmz * using size thresholds of >20kb for CNAs and 2Mb for cnloh unless the anomaly is in a recognized region of interest. All co-ordinates are hg36 and determined using OncoSNP except and (see below) Co-ordinates determined using Nexus (either due to variable CNA boundaries less well defined by OncoSNP or outside OncoSNP thresholds) called by OncoSNP as possible germline variant and manually verified as probable somatic event. Co-ordinates determined using Nexus Percentages represent CNA/cnLOH levels determined using OncoSNP. Percentages <10% and >90% are outside the OncoSNP thresholds. Approximate percentages derived using OncoSNP and visual inspection. NK= Not Known prior to testing; HRM= High Resolution Melting curve analysis; PFS= Progression Free Survival; W&W= Watch and Wait; CRS=Clinical Risk Score; N/A= Not Applicable; FU=Follow Up Flu(or F)=Fludarabine; Cy(orC)= Cyclophosphamide; M=Mitozantrone; minir=mini-rituximab 18

19 Supplementary Table 2 (cont). Clinical and Molecular Characteristics of B-CLL patients included in the study Patient No Age (yrs) Sex (M/F) IGHV gene status TP53 mutation status Current management PFS CRS CNAs and cnloh events * CLL M Unmutated Negative pre-chlorambucil <1year 2 46,XY, arr 11q22.3(104,158, ,733,134)x1,11q23.1q23.2(110,556, ,089,910)x1,11q23.3(chr11:115,527, ,182,933)x1, 13q14.3(49,038,530-50,906,239)x1, 17q21.33q22(45,720,750-49,017,380)x2 hmz CLL M Unmutated Negative pre-chlorambucil <1year 2 46,XY, arr 5q12.1(59,747,698-59,804,164)x3, 9q32(115,160, ,182,884)x3, 40% 12q12qter(37,473, ,088,962)x3, 80% 13q14.2q14.3 (47,443,974-50,755,648)x1, 80% 22q11.21(17,245,391-17,832,230)x1 CLL M Unmutated Negative pre-flu/cy <1year 2 47,XY, arr 90% 4q31.1(153,631, ,935,927)x1, 70% 12(0-132,146,663)x3, 17p12(13,118,349-13,171,201)x3 CLL M Mutated Cys135Ser pre-ofatumumab <1year 2 46,XY, arr 46,XY, arr 2q36.2 (225,094, ,617,560)x1, ~20% 9p23p21.1(11,986,163-32,324,220)x1, 50% 9q13qter(70,223, ,946,600)x3, 13q14.3(49,515,641-51,327,198)x0, 17pterp11.2(0-21,430,683)x1, 19q13.12q13.2(41,320,887-43,412,844)x2 hmz CLL M NK Negative pre-chlorambucil <1year 2 47,XY, arr 12q14.1(60,235,935-60,327,516)x1 CLL M NK Negative pre-flu/cy <1year 2 46,XY, arr 80% 5p13.3(37,408,772-37,784,135)x3, 80% 9pter9p11.2(0-43,254,687)x2 hmz, 70% 13q14.12q14.2(45,377,644-45,760,335)x3, 70% 13q14.2(46,403,477-47,574,445)x3, >90% 13q14.2(47,583,917-48,099,556)x1, 90% 13q14.2q14.3(48,719,245-50,523,975)x1, 70% 13q14.3qter(50,523, ,121,252)x3 CLL F Unmutated Negative pre-flu/cy < 1year 2 46, XX, arr 11q14.1q24.1(77,652, ,488,138)x1, 13q14.11(39,758,892-39,909,885)x0, 13q14.3(49,425,974-50,389,210)x0, 13qcenqter(19,075, ,142,980)x2 hmz CLL F Unmutated Negative pre-chlorambucil <1year 2 None noted CLL F Mutated Negative pre-chlorambucil 1year 2 46, XX, arr 13q (33,726,200-53,619,700)x1, 13q14.3(49,461,448-50,635,061)x0 CLL M Unmutated Negative pre-flu/cy <1year 2 46,XY, arr 4q34.1(172,446, ,482,835)x1 CLL M Mutated Negative pre-chlorambucil <1year 2 46,XY, arr 3p26.3p26.1(0-7,571,446)x1, >90% Xq27.1qter(139,443, ,582,606)x2 CLL087 NK M Unmutated Negative pre-chlorambucil <1year 2 47,XY, arr 80% 12(0-132,146,663)x3, Xp11.3p11.4(41,074,455-43,983,367)x0 CLL F Mutated Negative pre-flu/cy <1year 2 46,XX, arr >90% 13q14.12q14.3(45,817,763-49,547,499)x1 CLL M NK NK pre-chlorambucil <1year 2 47,XY, arr 80% 2q14.3qter(129,537, ,497,808)x2 hmz, 6q14.1q21(79,968, ,917,920)x1, 70% 12(0-132,146,663)x3, >90% 20q11.2qter (30,155,027-62,435,964)x2 hmz, 22q11.23(35,010,463-35,020,688)x0 CLL102 NK M Unmutated Negative NK <1year 2 47,XY, arr 80% 12(0-132,146,663)x3 CLL103 NK F Unmutated Negative NK <1year 2 47,XX, arr 1q25.3(183,360, ,397,169)x3, >90% 2q22.1q31.1(141,094, ,777,386)x2 hmz, >90% 9p24.1p22.3(9,000,336-14,313,346)x2 hmz, 10q26.3(135,089, ,374,737)x3, 80% 12(0-132,146,663)x3, >90% 13q33.3q34(108,839, ,014,298)x2 hmz, 20p11.23(19,723,376-19,823,725)x1, 70% 22q11.21q12(17,295,655-19,791,274)x3 CLL106 NK M Unmutated Arg267Pro NK <1year 2 46,XY, arr 90% 1p34.2(41,538,823-42,112,410)x1, 90% 6p25.3p25.1(1,376,135-6,173,214)x1, 90% 6p25.1p24.3(6,471,929-7,578,657)x1, 90% 6p24.3p24.1(10,155,458-11,223,334)x1, 90% 6p24.1p22.3(12,754,522-18,670,995)x1, 90% 6p21.2(38,889,727-39,544,026)x1, 90% 6p21.1p12.3(43,689,563-46,299,596)x1,90% 7p15.3(23,346,710-24,400,934)x1, 90% 9q33.3 (127,381, ,900,525)x1, 90% 12p13.33(977,044-1,959,491)x1, 17pterpcen(0-22,125,162)x1, 18pterp11.31(0-6,572,104)x1, 18p11.31p11.23(6,802,509-7,333,763)x3, 18p11.22p11.21(8,926,107-13,281,880)x1, 18p11.21(13,536,937-15,348,275)x1, 90% 22q11.21(16,712,510-17,055,914)x1, 90% 22q11.23q12.1(23,409,925-27,331,789)x1, 90% 22q12.1q12.2(27,573,410-27,918,140)x1 CLL M Unmutated Arg273His pre-fcr <1year 2 46,XY, arr <10% 3p25.1p14.3(12,516, )x1, <10% 5q14.3q23.3(87,285, ,855,075)x1, >90% 6p24.3p22.3(10,220,342-16,992,531)x2 hmz, 6q16.3(100,876, ,445,547)x1, 80% 8q22.3q24.21( , ,286,818)x2 hmz, 80% 8q24.21(128,286, ,297,901)x3, 80% 8q24.21-q24.3(128,297, ,274,826)x2 hmz, 12q14.1q14.2(56,925,673-62,754,801)x2 hmz, 16q22.3q23.1(72,290,299-76,525,557)x2 hmz, <10%-50% 17p13.3p11.2(0-18,485,225)x2 hmz, ~30% 18q11q23(22,045,087-76,115,172)x2 hmz, 80% 22q11.23(23,991,557-24,250,622)x1 CLL M NK Negative pre-alemtuzumab <1year 2 46,XY, arr 80% 5q23.3 (124,282, ,531,381)x3, ~10% 19q13.12q13.13(41,440,201-43,149,438)x1 hmz ARR M Unmutated Negative pre-bendamustine <1year 2 46,XY, 13q14.2q14.3(49,983,042-51,331,787)x1 CLL M Mutated Negative chemotherapy CLL F NK NK chemotherapy CLL M Unmutated Tyr205Cys chemotherapy CLL M Unmutated Negative chemotherapy CLL F Unmutated Pro153- AlafsX28 chemotherapy CLL M Mutated Glu271Lys chemotherapy * using size thresholds of >20kb for CNAs and 2Mb for cnloh unless the anomaly is in a recognized region of interest. All co-ordinates are hg36 and determined using OncoSNP except and (see below) N/A 3 46,XY, arr 9q32qter(114,777, ,273,252)x2 hmz, 13q14.3(49,448,465-49,958,586)x1, <50%13q14.3(49,958,586-50,470,735)x1 Co-ordinates determined using Nexus (either due to variable CNA boundaries less well defined by OncoSNP or outside OncoSNP thresholds) called by OncoSNP as possible germline variant and manually verified as probable somatic event. Co-ordinates determined using Nexus Percentages represent CNA/cnLOH levels determined using OncoSNP. Percentages <10% and >90% are outside the OncoSNP thresholds. Approximate percentages derived using OncoSNP and visual inspection. N/A 3 46,XX, arr 1p36.22(10,166,938-10,240,277)x3, 30% 3p25.3(9,610,471-10,363,762)x1, 30% 3p21.33p21.31(44,436,042-47,426,525)x1, 40% 4pterp14( )x1, 30-40% 5pterp13.3(0-33,225,747)x3, 40% 5q31.3q33.3(141,585, ,102,574)x1, 40% 5q33.3q34(159,368, ,909,332)x1, 6q16.1(94,130,248-94,158,093)x3, 70% 8q21.3qter(88,785, ,263,890)x3, <10% 12p13.2p13.1(10,176,350-14,248,326)x1; 80% 13q12.11q12.13(17,920,393-25,846,844)x1, 13q14.2q21.1(46,976,461-52,641,590)x1, ~10% 15q15.1(38,975,579-40,758,191)x1, 90% 17pterp11.2(0-22,684,316)x1, 80% 17q12(30,750,407-31,140,027)x1, 18pterp11.21(0-14,078,284)x1, <10% 18q11.2q21.2(20,999,500-49,134,000)x1, 90% 19p13.3(212,341-4,841,219)x1, 80% 19p13.3(4,841,219-5,635,144)x1, 70% 19p13.2(9,104,258-10,926,409)x3, <10% Xq12qter(66,869, ,913,750)x1 N/A 3 46,XY, arr ~30% 1p36.31p36.23(6,508,255-7,966,180)x3, 1p21.1(105,011, ,118,455)x3, 70% 2q36.3q37.1(230,780, ,367,190)x1, 90% 3pter3p12.2(0-83,717,142)x1, 90% 3p12.1pcen(83,723,698-90,587,335)x4, 80% 3qcen3qter(94,988, ,302,161)x3, 50% 7q36.1qter(151,287, ,803,229)x1, 90% 8pterp21.1(0-31,094,002)x1, 90% 8p12(33,488,001-36,433,467)x1, 90% 8p11.22p11.21(39,796,895-40,860,746)x1, 90% 8p11.21pcen(40,947,586-43,533,112)x1, 70% 11q13.4(73,346,490-74,926,833)x1, 80% 15q15.1q23(40,832,872-66,297,824)x3, 50% 15q23qter (66,297, ,176,830)x3 (partial tetrasomy?), 16q23.1(77,240,683-77,338,055)x1, 90% 17pterpcen(0-22,125,162)x1,70% 17qcenqter(22,496,341-78,503,256)x3, ~90% Xp22.33(0-2,709,234)x2 hmz, ~90% Xq21.31q21.32(83,356, ,208,942)x2 hmz, ~90% Xq28(154,570, ,913,754)x2 hmz N/A 3 46,XY, arr 5pterp15.1(0-15,968,582)x2 hmz, 13q14.12q14.2(45,353,865-46,421,078)x1, 13q14.2q14.3(47,387,659-50,923,001)x1, 13q21.2q33.3(58,195, ,914,455)x2 hmz, 20p12.1(12,586,831-15,008,138)x2 hmz N/A 3 46,XX, arr 12q24.31(119,578, ,172,890)x1, 13q12.2(26,955,745-27,052,913)x1, 13q12.2(27,304,454-27,383,456)x1, 13q12.2(27,461,395-27,710,289)x1, 13q12.3(27,878,812-28,049,348)x1, 13q14.2(46,713,975-47,895,694)x1, 13q14.2(47,895,694-48,559,998)x0, 13q14.2(48,559,998-48,704,022)x1, 13q14.2q14.3(48,704,022-50,191,807)x0, 13q14.3(50,191,807-50,302,346)x1, 70% 15q11.2(20,302,387-20,657,533)x3, 80% 15q14qter(33,100, ,176,830)x3, 90% 17pterp11.2(0-16,508,847)x1 N/A 3 46,XY, arr <10% 13q13.3q34(36,483, ,123,125)x1, 70% 13q14.2q14.3(48,713,406-50,479,598)x1, 13q14.3(49,229,284-50,423,376)x0, 30-70% 17pterp11.2(0-21,100,678)x2 hmz NK= Not Known prior to testing; PFS= Progression Free Survival; W&W= Watch and Wait; CRS=Clinical Risk Score; N/A= Not Applicable;FU=Follow Up Flu(or F)=Fludarabine; Cy(or C)= Cyclophosphamide; R=Rituximab 19

20 Supplementary Table 2 (cont). Clinical and Molecular Characteristics of B-CLL patients included in the study Patient No Age (yrs) Sex (M/F) IGHV gene status TP53 mutation status Current management CLL074 NK F NK NK chemotherapy CLL075 NK M NK HRM Positive chemotherapy CLL F NK Negative chemotherapy PFS CRS CNAs and cnloh events * N/A 3 47,XX, arr 80% 3(0-199,264,624)x3, 4q13.3(71,086,180-71,455,900)x1, 4q13.3(71,736,694-71,822,824)x1, 11p15.4(9,745,400-10,316,737)x1, 11p13(32,809,601-33,299,316)x1, 11p13(33,682,105-34,179,798)x1, 13q14.2qter(47,632, ,121,252)x2 hmz, 19p13.13p13.12(12,921,191-14,497,928)x1, 19p13.12p13.11(15,839,641-16,595,885)x1, 19p13.11(18,336,675-19,169,002)x1, 19q13.32(51,526,523-53,157,140)x1 N/A 3 46,XY, arr 50% 1p36.23p36.32(2,396,446-8,492,493)x1, 50% 1p35.1p35.3(28,122,270-32,692,291)x1, 50% 1p32.21p32.3(55,244,056-56,962,381)x1, 90% 1q24.2q24.3(167,889, ,636,321)x2 hmz, 50% 1q32.2q41(212,087, ,475,787)x1,<10% 2p22.2p21(37,413,860-43,848,439)x1, 80% 9p24.19p13.3(5,944,274-33,693,234)x1, 80% 9q13q21.31(70,174,192-80,333,445)x1, 80% 9q22.33q32(101,069, ,321,935)x1, 80% 9q33.1(119,325, ,100,178)x1, 80% 9q33.1q33.2(121,461, ,566,552)x1, 80% 9q34.13q34.3(134,588, ,101,448)x1, 80% 13q14.12q14.2(45,128,861-47,516,428)x1, 90% 13q14.2q14.3(47,516,428-49,482,152)x1, 13q14.3 (49,482,152-49,732,093)x0, 90% 13q14.3(49,732,09-49,941,789)x1, 13q14.3 (49,941,789-50,302,346)x0, 90% 13q14.3(50,302,346-50,504,198)x1, 30% 14q21.3qter(47,690, ,581,072)x2 hmz, 70% 17pterp13.1(0-8,074,364)x2 hmz, Xq27.3 (142,536, ,254,333)x0 N/A 3 47,XX, arr ~80% 1p36.12p36.11(21,423,975-23,960,71)x2 hmz, ~50% 1pterp36.12(0-21,423,975)x2 hmz, 2q37.1q37.1(230,940, ,540,187)x2 hmz, 80%12(0-132,146,663)x3 CLL M Unmutated Negative chemotherapy CLL F Unmutated Negative chemotherapy CLL M Unmutated Cys135Ser chemotherapy CLL M Unmutated Negative chemotherapy CLL M Unmutated Negative chemotherapy CLL F NK Negative chemotherapy CLL M NK Negative chemotherapy N/A 3 46,XY, arr 90% 20q11.21q13.12(31,081,079-43,771,286)x2 hmz, arr 20q13.12(43,780,732-43,812,931)x0, 90% 20q13.12,qter(43,812,931-62,385,675)x2 hmz N/A 3 46,XX, arr 60% 4q13.1q22.2(59,939,995-95,374,528)x1, 60% 8p22p21.3(16,324,942-20,039,997)x1, 11q22.1(100,699, ,935,421)x1, 11q22.3q23.2(107,417, ,576,024)x1, 90% Xq28(153,861, ,095,678)x0 <1year 3 46,XY, arr <10% 2pterp14(0-67,981,586)x3, 90% 2q11.1q11.2(95,597,687-97,676,314)x1, 90% 2q14.2q14.3(121,572, ,531,524)x1, 90% 2q14.3(127,036, ,569,251)x1, 90% 2q21.2q24.1(134,752, ,542,922)x1, 40% 8q22.2qter(100,925, ,263,890)x3, 90% 9pterp11.2(0-33,318,812)x1, 90% 9p13.3p11.2(33,831,777-43,548,708)x1, 9q21.13(74,372,294-74,516,002)x1, <10% 10q23.2q23.31(89,306,467-93,720,035)x1, 90% 11q12.2q13.1(61,231,809-64,790,450)x1, 70% 12q13.11q13.12(47,308,669-48,029,106)x3, 50% 13q11q13.1(17,920,393-32,659,455)x1, 50% 13q14.3(48,898,975-50,805,219)x1, 90% 17pterpcen(0-22,125,162)x1, 70% 17qcenqter(22,713,949-78,637,198)x3; 90% 19pterp13.2(0-10,071,486)x1; 70% 19p13.2p13.13(10,071,486-13,419,854)x3 <1year 3 46,XY, arr 2p16.1(55,330,268-55,352,382)x1, 11q22.3q24.3(103,008, ,852,266)x1, 13q14.3(49,500,283-50,255,794)x1, ~20% 20q11.22qter(32,772,470-62,385,675)x2 hmz <1year 3 46,XY, arr 30% 2q33.2q36.3(206,690, ,780,953)x1, 40% 17pterp11.2(0-20,067,160)x1, 60% 17p11.2(20,067,161-21,471,049)x1 N/A 3 46,XX, arr 60% 2pterp16.1(0-60,616,006)x3, 4pterp14(0-37,269,568)x1, 4p13q12(41,367,356-58,152,039)x1, 4q13.1q13.2(64,455,242-66,975,683)x1, 4q13.2q13.3(69,847,111-71,783,789)x1, 60% 8pterq12.1(0-63,281,507)x1, 11q21q24.2(92,396, ,058,859)x1, 60% 18pterp11.21(0-12,950,950)x1, 40% 18p11.21(12,950,950-14,856,930)x1, N/A 3 46,XYarr 90% 2q35(218,476, ,204,673)x1, 90% 2q36.3q37.1(230,239, ,446,513)x1, 8q21.11q21.13(78,181,345-80,471,188)x2 hmz, >90% 13q11.2qter(20,381, ,908,097)x2 hmz, 13q14.3(49,494,577-50,238,316)x0, 80% 16p12.1(21,860,290-22,350,261)x1 ARR M Unmutated Negative chemotherapy MCLL M NK Ile195del chemotherapy N/A 3 46,XY, arr 50% 2p16.1p14(60,279,157-67,896,939)x3, 70% 15q25.2q25.3(82,852,550-83,573,559)x3,~10% 18q23(75,210,150-76,116,152)x1, 20q11.23(34,645,412-34,974,353)x1 <6 months 3 46,XY, arr 2p16.1(56,456,704-56,519,123)x1, 50% 4q21.21-q21.23(82,819,454-84,544,925)x1, 30% 5q21.3q22.1(106,355, ,412,993)x1, 30% 5q22.2q22.3(112,371, ,463,403)x1, 30% 11q22.1q22.2(99,750, ,130,166)x1, 30% 11q22.3q23.2(104,918, ,505,937)x1, 30% 11q23.2q23.3(114,934, ,687,486)x1,30%11q23.3(117,457, ,510,342)x1, 30% 11q23.3(119,441, ,562,032)x1, 30% 11q24.2q24.3(126,353, ,613,424)x1, 30% 11q24.3(128,805, ,951,253)x1, 50% 12q24.11(109,411, ,658,160)x1, ~10% 13q13.3q14.3(36,320,198-49, )x1, ~90% 13q14.3(49, ,493,361)x1, 13q14.3(49,493,362-49,639,921)x0, ~90% 13q14.3(49,639,922-49,845,114)x1, 13q14.3(49,845,115-49,992,953)x0, ~90% 13q14.3(49,992,954-50,209,795)x1, 13q14.3(50,209,796-50,309,646)x0, ~10% 13q14.3(50,309,647-50,383,609)x1, ~10% 13q14.3q34(50,383, ,125,255)x1, 60% 17p13.3-p11.2(0-17,522,348)x2 hmz, Xp11.4p11.3(41,001,474-43,480,372)x0 CLL039 NK F NK Negative NK NK NK 46,XX, arr 4q24(107,162, ,237,623)x3, 40% 10p12.33(17,900,269-18,398,113)x1, 40% 10q24.32q25.2 (103,564, ,658,371)x1, ~10% 16q23.1(74,997,818-75,060,214)x1 CLL M Unmutated Negative pre-flu/cy FU<1year NK None noted CLL M NK Negative pre-flu/cyr FU<1year NK 46,XY, arr 70% 11q14.1q23.3(77,741, ,465,523)x1 CLL M NK HRM Positive pre-alemtuzumab FU<1year NK 46,XY, arr 80% 17pterp11.2(0-21,964,965)x1, 80% 20p13-p12.2(4,785,063-9,654,363)x1, 80% 20p12.1q11.21(17,106,967-29,434,300)x1, 80% 20q11.22q13.12 (32,667,050-42,799,625)x1 CLL M NK NK pre-br FU<1year NK 46,XY, arr 12p13.33(1,302,036-1,460,642)x3, 16p13.3(3,601,816-3,736,149)x1, 90%16q24.3(87,195,552-88,104,547)x1, 17q21.2(37,362,610-37,545,814)x1 CLL M Unmutated Negative pre-flu/cyr FU<1year NK 46,XY, arr 70% 1p34.1(46,022,981-46,260,143)x3, 90% 1p22.2(88,981,462-89,083,655)x1, 90% 1p22.2(89,319,559-89,741,569)x1, 90% 1p13.1(116,804, ,487,515)x1, 2q34(214,363, ,512,575)x1, 2q34(214,639, ,706,974)x1, 90% 10q11.23(50,320,838-50,505,621)x1, 90% 10q11.23(51,479,606-52,200,825)x1, 11q22.3(107,443, ,149,178)x1, 12q24.31(121,703, ,806,600)x1, 90% 13q14.11(40,332,062-40,718,960)x1, 90% 13q14.11(41,349,414-42,201,691)x1, 90% 13q14.2q14.3(47,731,453-51,632,859)x1 CLL M NK Negative pre-flu/cyr FU<1year NK 46,XY, arr 90% 5q14.3q23.1(88,216, ,359,270)x1, 9q22.31(94,191,198-94,271,909)x1, 90% 11q22.3q23.3(103,276, ,197,768)x1 CLL M NK Arg267Pro pre-flu/cyr FU<1year NK 46,XY, arr 30% 13q14.2q14.3(47,411,558-50,943,647)x1 MCLL M NK Negative pre-chlr FU<1year NK 47,XY, arr 70% 12(0-132,146,663)x3 MCLL F NK Negative pre-chlorambucil FU<1year NK 46,XX, arr 70% 12(0-132,146,663)x3, 70% 17q22(51,452,730-51,748,539)x3 MCLL137 NK M NK NK NK NK NK 46,XY, arr 1q24.2q43(230,738, ,657,439)x2 hmz, 2q35q36.1(220,876, ,498,801)x1, 13q14.2q14.3(48,748,895-51,339,176)x1 MCLL138 MCLL M M Mutated NK Negative Negative pre-chlo pre-chlr FU<1year FU<1year NK NK 46,XY, arr 1q25.3q31.1(183,714, ,107,919)x2 hmz, 50% 9q21.11q34.3(71,087, ,233,916)x2 hmz, 80% 13q14.11q21.33(40,811,871-45,706,420)x1, ~90% 13q14.12q14.3(45,706,421-49,971,647)x1, 46,XY, arr 2q11.2q12.1(100,630, ,604,613 >90% 13q14.3(49,971,648-51,240,377)x1, )x2 hmz, 40% 13q12.11q34(19,727,582-80% 114,125,259)x2 hmz, 80% 13q14.3(49,443,358-50,378,319)x1, 17q11.2(24,916,898-24,975,608)x3 hmz MCLL M Mutated Negative pre-br FU<1year NK 46,XY, arr 15q26(100,010, ,338,915)x1, 17q11.2(24,913,908-24,975,608)x3 MCLL F Mutated Negative pre-fcm-minir FU<1year NK 46,XX, arr 80% 13q14.3(49,552,639-49,967,095)x1, 18q23(72,560,419-72,606,933)x1 MCLL M Mutated Negative pre-br FU<1year NK 46,XY, arr 13q14.3(49,428,032-50,557,052)x1, 13q14.3(49,516,595-50,2557,94)x0 MCLL M Unmutated Negative pre-fcr FU<1year NK 46,XY, arr ~10% 4pterp15.1(0-28,404,856)x1, ~10% 11q14.3q23.3(90,803, ,404,675)x1, 20% 13q14.2q(46,483,567-47,832,800)x3, 20%13q14.2q21.33(46,483,568-70,853,744)x2 hmz(or x1), 15q25.3(83,260,487-83,303,854)x1, 20q11.23(34,428,530-35,199,710)x2 hmz, * using size thresholds of >20kb for CNAs and 2Mb for cnloh unless the anomaly is in a recognized region of interest. All co-ordinates are hg36 and determined using OncoSNP except and (see below) Co-ordinates determined using Nexus (either due to variable CNA boundaries less well defined by OncoSNP or outside OncoSNP thresholds) called by OncoSNP as possible germline variant and manually verified as probable somatic event. Co-ordinates determined using Nexus Percentages represent CNA/cnLOH levels determined using OncoSNP. Percentages <10% and >90% are outside the OncoSNP thresholds. Approximate percentages derived using OncoSNP and visual inspection. NK= Not Known prior to testing; PFS= Progression Free Survival; W&W= Watch and Wait; CRS=Clinical Risk Score; N/A= Not Applicable;FU=Follow Up Flu(or F)=Fludarabine; Cy(or C)= Cyclophosphamide; Chl=Chlorambucil; R=Rituximab; B=Bendamustine; O-Ofatumumab; M=Mitozantrone; minir=mini-rituximab 20

21 Supplementary Table 3. Validation of OncoSNP Quantification by comparison with FISH data Sample ID a OncoSNP Result b FISH OncoSNP vs FISH Correlation (Yes/No) c CLL041 homozygous del13q % homozygous 10% heterozygous del13q14.3 Yes CLL060 80% heterozygous del13q % heterozygous del13q14.3 Yes CLL071 homozygous del13q % homozygous del13q14.3 No CLL058 90% heterozygous del17p % heterozygous del17p13.1 Yes CLL080 50% heterozygous del13q % homozygous del13q14.3 Yes CLL081 90% heterozygous del17p % heterozygous del17p13.1 Yes 50% heterozygous del13q % heterozygous del13q14.3 Yes CLL082 homozygous del13q % homozygous del13q14.3 Yes CLL084 No change noted 15% heterozygous del13q14.3 No CLL085 50% heterozygous del13q % heterozygous del13q14.3 Yes CLL087 80% trisomy 12 82% trisomy 12 Yes CLL090 80% heterozygous del13q % heterozygous del13q14.3 Yes 80% trisomy 12 66% trisomy 12 Yes CLL092 90% heterozygous del11q % heterozygous del11q22.3 Yes 90% heterozygous del13q % heterozygous del13q14.3 Yes CLL092 (relapse) 90% heterozygous del11q % heterozygous del11q22.3 Yes 90% heterozygous del13q % heterozygous del13q14.3 Yes CLL093 90% heterozygous del13q % heterozygous del13q14.3 Yes CLL094 40% heterozygous del17p % heterozygous del17p13.1 Yes CLL095 70% trisomy 12 57% trisomy 12 Yes a All samples are pretreatment unless otherwise stated b Mean OncoSNP % is given where segmentation of the data results in multiple calls across a region (eg trisomy 12 examples) In the case of homozygous losses, OncoSNP percentages of 100% are not stated in the Table as formal calibration testing for homozyous events is pending c Correlation was defined as less than 15% difference between FISH and OncoSNP 21

22 Supplementary Table 4. Statistical analyses of IgVH SHM status vs CRS, Number of CNAs/cnLOHs and Length of CNAs/cnLOHs IGHV gene status Unmutated Mutated Covariate Category n % 1 n % 1 P-value 2 CRS 0: W&W : PFS>1year : PFS<1year : Chemotherapy All Total number of CNAs/cnLOHs > All Total length of CNAs/cnLOHs <1Mb Mb >5Mb All Percentage calculated over the total number of patients for whom data were available 2 Cochran-Armitage trend exact test Supplementary Table 5. Statistical analyses of TP53 mutation status vs CRS, Number of CNAs/cnLOHs and Length of of CNAs/cnLOHs TP53 mutation status Negative Positive Covariate Category n % 1 n % 1 P-value 2 CRS 0: W&W : PFS>1year : PFS<1year : Chemotherapy All Total number of CNAs/cnLOHs > All Total length of CNAs/cnLOHs <1Mb Mb >5Mb All Percentage calculated over the total number of patients for whom data were available 2 Cochran-Armitage trend exact test 22

23 Supplementary Table 6. Detailed results for patients with 13q14.3 abnormalities Patient No Chr13 abnormalities % Cells with anomaly encompassing 13q14.3 MDR Additional abnormalities (Pre-treatment)* Clinical details PFS Clinical risk score RB1 deletion or cnloh (Y/N) MDR only CLL037 del13q14.3 (MDR) 100 cnloh 9q32-q34.3 chemotherapy N/A 3 N CLL090 80% del13q14.3 (MDR) 80 90% del7q33-q34; 50% cnloh11q12.1-qter chemotherapy N/A 3 N (ATM ); 80% tri12 CLL092 del13q14.3 (MDR) 100 del2p16.1(<500kb); del11q22.3-q24.3 (ATM ); chemotherapy N/A 3 N ~20%cnLOH20q11.22-qter CLL110 homozygous del13q14.3 (MDR) 100 None noted chemotherapy N/A 3 N MCLL100 80% del13q del18q23(<500kb) pre-fcm-minir FU<1year NK N More than MDR N/A CLL005 80% del13q14.2-q ampl5q12.1 (<500kb); 40% ampl12q12-qter; 80% W&W N/A 0 Y del22q11.21 CLL034 del13q14.2-q % ampl2p14-pter; del11q22.3 (ATM ) W&W N/A 0 N CLL041 del13q14.12-q14.3 with homozygous del13q cnloh6p ; 80% del10q24.32 W&W N/A 0 Y q14.3) CLL053 cnloh13q12.11-qter with homozygous del13q None noted W&W N/A 0 N CLL060 50% del13q14.12; <10% del13q14.2; 50% del13q14.2; 80% del13q del1p33(<500kb) ampl7p22.1(<500kb), ampl11p13; <10% del 16p13.3 W&W N/A 0 Y CLL064 50% del13q14.2-q14.3; 40% del13q ampl3q13.32(<500kb) W&W N/A 0 Y CLL070 del13q14.2-q None noted W&W N/A 0 N CLL080 30% del13q14.11-q14; ~50% del13q del3q25.3(<500kb) pre-chl >1year 1 N CLL085 50% del13q14.11-q cnloh10q23.1-q23.31; pre-fc >1year 1 Y CLL093 del13q14.2-q14.3; % ampl6p ; del6q14.1-qter; 70% pre-chl >1year 1 Y CLL096 40% cnlohchr13 with homozygous del13q ampl10q25.1(<500kb); 70% ampl20q13.13 (<500kb) None noted pre-fc >1year 1 N CLL097 >90% del13q14.11-q14.3 >90 None noted >1year 1 Y CLL % del13q14.2-q21.32; 50% ampl13q del11q14.1-qter (ATM ); 70% ampl22q11.21-qter >1year 1 Y q33.1 CLL105 90% del13q14.2-q14.3 with homozygous 90 80% del10q ; 90% del11q14.1-q24.2 NK >1year 1 Y del13q14.3 (ATM ) CLL107 >90% del13q14.11(two regions <500kb);>90% del13q14.2-q14.3 >90 50% del3p21.31;50% del 3p21.1, >90% del11q14.3-q24.3 (ATM ) >1year 1 Y CLL003 del13q del11q22.3 (ATM ); del11q23.1-q23.2; del11q23.3; pre-fc <1year 2 N cnlohq17q21.33-q22 CLL046 del13q14.12-q14.2; cnloh13q21.2-q33.3; 90 cnloh5pterp15.1; cnloh20p12.1 pre-chl <1year 2 Y del13q14.2-q14.3 CLL056 70% ampl13q14.12-q14.2; >90% del 13q14.2; 90% del13q14.2-q14.3; 70% ampl13q qter 90 80% 5p13.3 (<500kb); 80% cnloh 9p11.2-9pter pre-chl <1year 2 Y CLL063 cnloh13q12.11-qter with homozygous del13q del11q (ATM ) pre-fc <1year 2 N CLL082 del13q with homozygous del13q None noted pre-chl <1year 2 Y CLL088 >90% del13q14.3 >90 None noted pre-fc <1year 2 Y CLL100 del13q14.12-q del6q23.3(<500kb); del11q21-q23.3 (ATM ) NK <1year 2 Y ARR003 del13q14.2q None noted pre-b <1year 2 N CLL040 del13q14.2-q ampl1p36.22 (<500kb); 20% del3p21.31-p21.33; 20% del3p25.3; 40% del4p14-pter; 30-40% ampl5p13.3-pter; 40% del5q31.3-q33.3; 40% 5q33.3-q34, ampl6q16.1(<500kb), 70% ampl6q22.2(<500kb); ampl 8q21.3-qter; <10% del12p13.1-p13.2; 80% del13q12.11-q12.13; ~10% del15q.15.1, 90% del17p11.2-pter (TP53 ); del17q12(<500kb), del18p11.21-pter; <10% del18q12.2-q21.2; 90% del19p13.3; 80% del19p13.3, 70% ampl19p13.2;; <10% delxq12-qter CLL044 homozygous del13q del2q36.2, 20% del9p21.1-p23; 50% ampl9q13- qter; del17pcen-pter (TP53 ); cnloh19q q13.2 CLL050 del13q14.2-q del6q14.1 (<500kb) intermittent; ampl7p21.1 (<500kb), del11q12.3 (<500kb) CLL058 CLL071 del13q12.3(<500kb), del13q12.2(<500kb - two regions), del13q14.2-q14.3; homozygous del13q14.2-q14.3 (intermittent) <10% del<10% 13q13.3q34, 70% del13q14.2q14.3 with homozygous del13q14.3 (MDR) 100 del12q24.31; 80% ampl15q14-qter; 90% del17p11.2-pter (TP53 ), 70% ampl15q11.2 chemotherapy chemotherapy chemotherapy chemotherapy %-70% cnloh17p11.2-pter (TP53 ) chemotherapy CLL074 cnloh13q14.2-qter % tri3; del11p15.4, del19p13.12-p13.13; del19p13.11-p13.12, del19p13.11; del19q13.32 CLL075 80% del13q14.12-q14.2; 90% del13q14.2-q14.3 intermittent with homozygous del13q % del1p36.32-p36.23; 50% del1p35.3-p35.1; 50% del1p32.3-p32.2; 90% cnloh1q24.2-q24.3; 50% del1q32.2-q41; <10% del2p21-p22.2; high ampl6q23.3(<500kb); 80% del9p13.3-9p24.1; 80% del9q13-q21.31; 80% del9q22.33-q32; 80% del9q33.1; 80% del9q33.1-q33.2; 80% del9q q34.3; 30% cnloh14q qter; 70% cnloh17p13.1-pter (TP53 );delxq27.3 CLL081 50% del13q12.11-q13.1; 50% del13q14.3 (MDR) 50 10% ampl2p14-pter; 90% del2q ; 90% del2q14.2-q14.3, 90% del2q14.3; 90% del2q21.2- q24.1; 40% ampl8q22.2-qter; 90% del9p11.1-pter; del9q21.13(<500kb); <10% del10q10q23.2q23.31; 90% del11q12.2-q13.1; 70% ampl12q13.12; 90% del17pcen-pter (TP53 ); 70% ampl17qcen-qter; 90% del19p13.2-pter; 70% ampl19p13.13-p13.2 CLL112 MCLL148 >90% cnloh13q12.11-qter with homozygous del13q14.3 ~90% del13q14.3 with intermittent homozygous del13q14.3; ~10% del13q14.3q % del2q35; 90% del2q36.3-q37.1; cnloh 8q21.11q21.13, 80% del16p12.1 chemotherapy chemotherapy chemotherapy chemotherapy N/A 3 Y N/A 3 N N/A 3 N N/A 3 Y N/A 3 N N/A 3 N N/A 3 Y N/A 3 N N/A 3 N ~90 del2p16.1(<500kb); 50% del4q21.21-q21.23; 30% chemotherapy <6 months 3 N del5q21.3q22.1; 30% del5q22.2q22.3; 30% del11q22.1q22.2; 30% del11q22.3q23.2 (ATM ); 30% del11q23.2q23.3; 30% del11q23.3; 30% del11q23.3; 30% del11q24.2q24.3; 30% del11q24.3 (<500kb); 50% 12q24.11(<500kb); 60% cnloh17p13.3-p11.2 (TP53 ); delxp11.4p11.3(41;001;474-43;480;372) CLL068 del13q % del2q pre-fc NK NK Y CLL072 del13q14.2-q high ampl15q24.1(<500kb) NK NK NK N CLL118 90% del13q14.11 (two regions); 90% del13q14.2- q % ampl1p34.1(<500kb); 90% del1p22.2(two regoins <500kb); del2q34(two regions <500kb); del1p13.1; 90% del10q11.23(<500kb); 90% del10q11.23, del12q24.3(<500kb); del11q22.3 (ATM ) pre-fcr NK NK Y CLL120 30% del13q14.2q None noted pre-flu/cy-r FU<1year NK Y MCLL137 del 13q14.2q cnloh1q24.2q43; del2q35q36.1 NK NK NK N MCLL138 80% del 13q14,11q21.33, ~90% del 13q14.12q14.3, >90% del13q14.3 >9 cnloh 1q25.3q31.1; amp9q21.11q34.3; ampl16p11.2(<500kb) pre-chlo FU<1year NK Y MCLL169 40% cnloh13q12.11q34 with 80% del13q cnloh2q11.2q12; ampl17q11.2(<500kb) pre-chlr FU<1year NK N MCLL152 del13q14.3 with homozygous del13q None noted pre-br FU<1year NK N MCLL161 20% ampl13q14.2; 20% cnloh/del 13q14,2q ~10% del4pterp15.1, ~10% del11q14.3q23.3, 20% ampl13q14.2, del15q25.3 (<500kb), 20% cnloh20q11.23 * deletions events are heterozygous unless otherwise stated. Percentages of 30-90% arefrom OncoSNP; all other calls are from visual interpretation. CNAs are >500kb and cnloh>2mb unless otherwise stated; known regions of interest are highlighted in bold MDR = Minimally Deleted Region; PFS = Progression Free Survival; RB1 - Retinoblastoma gene NK=Not Known; Flu(or F)=Fludarabine; Cy(or C)= Cyclophosphamide; R= Rituximab; minir=mini-rituximab B=Bendamustine; Chl=Chlorambucil; O=Ofatumumab pre-fcr FU<1year NK Y 23

24 Supplementary Table 7. Statistical analyses of RB1 deletions or cnloh vs CRS and IGHV gene status in patients with 13q14.3 deletions RB1 deletion or cnloh No Yes Covariate Category n % 1 n % 1 P-value CRS 0: W&W : PFS>1year : PFS<1year : Chemotherapy All IGHV gene status Unmutated Mutated All Percentage calculated over the total of patients for which the data were available 2 Cochran-Armitage trend exact test 3 Chi-square test Supplementary Table 8. Statistical analyses of Clinical Risk Score vs Total Number and Total Length of CNAs/cnLOHs excluding and including well-known abnormalities Clinical Risk Score 0 (W&W) 1 (PFS>1year) 2 (PFS<1year) 3 (Chemotherapy Refractory) Covariate Category n % 1 n % 1 n % 1 n % 1 P-value 2 Total number of CNAs/cnLOHs excluding known abnormalities Total length of CNAs/cnLOHs excluding known abnormalities Total number of CNAs/cnLOHs including known abnormalities > All <1Mb Mb >5Mb All > All Percentage calculated over the total number of patients for whom data were available 2 Cochran-Armitage trend exact test Supplementary Table 9. Statistical analyses of Clonal Evolution vs CRS, Number of CNAs/cnLOHs and Length of of CNAs/cnLOHs Clonal Evolution No Yes Covariate Category n % 1 n % 1 P-value 2 CRS 0: W&W : PFS>1year : PFS<1year : Chemotherapy All Total number of CNAs/cnLOHs > All Total length of CNAs/cnLOHs <1Mb Mb >5Mb All Percentage calculated over the total number of patients for whom data were available 2 Cochran-Armitage trend exact test 24

25 Supplementary Table 10. KEGG Pathway Analysis KEGG ID Pathway* P-value Odds Ratio Exp Count Count Size Genes 4742 Taste transduction ADCY6,GNG3,KCNB1,TAS2R7,TAS2R8,TAS2R Phagosome COLEC12,COMP,MRC1,MRC1L1,OLR1,TLR4,TUBA1 A,TUBA1B,TUBA1C,TUBB2B,VAMP Pathogenic Escherichia coli infection ARPC4,TLR4,TUBA1A,TUBA1B,TUBA1C,TUBB2B 4210 Apoptosis ATM,BAD,CAPN1,CHP,IKBKB,IRAK2,TP Amyotrophic lateral sclerosis (ALS) BAD,CHP,GRIN2B,PRPH,TP Caffeine metabolism NAT1,NAT Small cell lung cancer CDKN1B,COL4A1,COL4A2,IKBKB,RB1,TP Salivary secretion ADCY6,ATP2B2,HTN1,MUC7,PLCB3,STATH 5215 Prostate cancer BAD,CDKN1B,CREBBP,IKBKB,RB1,TP Gap junction ADCY6,PLCB3,TUBA1A,TUBA1B,TUBA1C,TUBB2B 4720 Long-term potentiation CHP,CREBBP,GRIN2B,PLCB3,PPP1CC 5212 Pancreatic cancer BAD,IKBKB,RALGDS,RB1,TP Chronic myeloid leukemia BAD,CDKN1B,IKBKB,RB1,TP Malaria COMP,ICAM1,KLRK1,TLR Wnt signaling pathway CHP,CREBBP,DKK4,LRP6,PLCB3,PPP2R5B,SFRP1, TP Sulfur metabolism PAPSS2,SULT1E Cell cycle ANAPC5,ATM,CDKN1B,CDKN2D,CREBBP,RB1,TP Base excision repair FEN1,OGG1,POLB 4140 Regulation of autophagy ATG4D,ATG5,GABARAPL Ubiquitin mediated proteolysis ANAPC5,CUL3,KEAP1,SYVN1,UBE2NL,UBE3C,VHL 4114 Oocyte meiosis ADCY6,ANAPC5,CHP,PPP1CC,PPP2R5B,SPDYC 5200 Pathways in cancer BAD,CDKN1B,COL4A1,COL4A2,CREBBP,ETS1,IKBK 592 alpha-linolenic acid metabolism B,NFKB2,RALGDS,RB1,SUFU,TP53,VHL FADS2,PLA2G4E 860 Porphyrin and chlorophyll metabolism FTH1,HMOX2,UGT2A Biosynthesis of unsaturated fatty acids FADS1,FADS Dorso-ventral axis formation ETS1,ETV Amoebiasis COL4A1,COL4A2,COL5A1,PLCB3,TLR4 983 Drug metabolism - other enzymes NAT1,NAT2,UGT2A Non-small cell lung cancer BAD,RB1,TP Ubiquinone and other terpenoidquinone HPD biosynthesis 140 Steroid hormone biosynthesis CYP17A1,SULT1E1,UGT2A Calcium signaling pathway ATP2B2,CHP,P2RX4,P2RX7,PLCB3,TNNC1,VDAC ECM-receptor interaction COL4A1,COL4A2,COL5A1,COMP 5131 Shigellosis ARPC4,ATG5,IKBKB 5210 Colorectal cancer BAD,RALGDS,TP African trypanosomiasis ICAM1,PLCB Neurotrophin signaling pathway BAD,IKBKB,IRAK2,RPS6KA4,TP SNARE interactions in vesicular SNAP23,VAMP3 transport 4360 Axon guidance CHP,EFNA5,EPHA5,RND1,UNC5D 4920 Adipocytokine signaling pathway ADIPOR2,CAMKK2,IKBKB 5145 Toxoplasmosis BAD,IKBKB,PLA2G4E,TLR4,TYK Focal adhesion BAD,COL4A1,COL4A2,COL5A1,COMP,PPP1CC,VAV 4115 p53 signaling pathway ATM,SESN1,TP Pancreatic secretion ADCY6,ATP2B2,PLA2G4E,PLCB Alzheimer's disease BAD,CAPN1,CHP,COX8A,GRIN2B,PLCB Renal cell carcinoma CREBBP,ETS1,VHL 4650 Natural killer cell mediated cytotoxicity CHP,ICAM1,KLRD1,KLRK1,VAV RNA degradation CNOT7,DCP1B,DCP Melanoma BAD,RB1,TP Tyrosine metabolism DBH,HPD 4520 Adherens junction CREBBP,FER,YES Bladder cancer RB1,TP Gastric acid secretion ADCY6,KCNJ1,PLCB3 603 Glycosphingolipid biosynthesis - globo GBGT1 series 3450 Non-homologous end-joining FEN B cell receptor signaling pathway CHP,IKBKB,VAV VEGF signaling pathway BAD,CHP,PLA2G4E 310 Lysine degradation SETD1B,SETD Cardiac muscle contraction CACNA2D4,COX8A,TNNC Huntington's disease COX8A,CREBBP,GRIN2B,PLCB3,TP53,VDAC3 564 Glycerophospholipid metabolism AGPAT6,DGKH,PLA2G4E 4330 Notch signaling pathway CREBBP,DLL Protein digestion and absorption COL4A1,COL4A2,COL5A Vascular smooth muscle contraction ADCY6,PLA2G4E,PLCB3,PPP1CC 360 Phenylalanine metabolism HPD 450 Selenocompound metabolism PAPSS mrna surveillance pathway MAGOHB,PPP2R5B,UPF1 561 Glycerolipid metabolism AGPAT6,DGKH 100 Steroid biosynthesis TM7SF Endometrial cancer BAD,TP53 * 'Top' 70 pathways only listed 25

26 B-CLL Study patients (n=93) B-CLL Study samples (93+42=135) Paired Pre-treatment and Relapse samples (n=42 pairs) Pre-treatment samples only (n=93) All samples investigated for intra-clonal heterogeneity using arrays and OncoSNP IGHV gene mutation status (n=67) and TP53 mutation status (n=83) established All Pre-treatment and Relapse pairs investigated for clonal evolution over time using arrays and OncoSNP Supplementary Figure 1. Flowchart documenting numbers of samples and overview of study. 26

27 Supplementary Figure 2. The Vysis LSI ATM FISH probe (boundaries indicated by yellow arrow and dashed lines) binds to non-deleted regions in two B-CLL patients with small 11q22.3 MDR deletions (indicated by red bars). 27

28 a) cnloh detected by Illumina chip b) High resolution melting curve c) Sanger sequencing of TP53 Wild-type TP53 Fluorescence intensity ratio CLL071 B-allele frequency Mutated TP53 Supplementary Figure 3. TP53 Mutation screening of the sample from CLL071 with cnloh encompassing this gene. a) Nexus image showing cnloh across the region b) high resolution melting curve analysis shows deviation from the TP53 reference sequence c) Sanger sequencing confirms the presence of a GAG>AAG mutation in this patient sample. 28

29 B-allele frequency a) Normal ~10% ~20% 30% 35% 40% 45% 50% 55% 60% 70% 75% 80% 85% 90% 95% 100% Manual OncoSNP Manual b) Log fluorescence ratio B-allele frequency Chromosome 6 position (Mb) Supplementary Figure 4: a) Nexus image showing examples of B-allele frequency plots for mosaic copy number losses called by OncoSNP and called manually. Here, the OncoSNP output percentages are given to the nearest 5%; however, for accuracy, the percentages in the data tables are rounded to the nearest 10% (1 significant figure). Currently, OncoSNP is considered accurate for calls between 10-90%, though this depends ultimately on the noise variances of the SNP data. Additional calls were made when mosaic changes were evident on visual inspection of plots b) Example of Nexus log fluorescence ratio plot and B-allele frequency plot for CLL084_R, demonstrating putative <10%6pterp22.3 loss that was not identified by OncoSNP. Region of loss is shown by the red double-ended arrow. 29

30 47 B-CLL patients with deletions/cnloh involving 13q13.4 Chromosome % 50% 0% 50% 100% Genes Exons CNVs mirna MDR (~7kb) involves DLEU2 Supplementary Figure 5. Nexus (BioDiscovery) image showing Illumina 1M- Duo data calls from 47 patients with deletions/cnloh involving 13q13.4. Red bar = CNA loss (thick bar indicates homozygous loss); green bar = CNA gain; mustard bar = LOH; purple bar = allelic imbalance.. 30

31 Supplementary Figure 6. a) The B-allele frequency plot shows clear evidence of mosaicism involving chromosome 13q14.1-q14.3 (red arrow) in a pre-treatment sample from CLL080 b) clonal evolution of a chromosome 13q14.3 deletion in a pre-treatment sample from patient CLL071 - the red, dashed arrows indicate the apparent boundary extension involving one allele. 31

32 Leukemic Progenitor(s) Pre-treatment Relapse CNAs/cnLOHs identified 2p16.1(x1) 11q22.1(x1) 11q22.3q24.3(x1) 13q14.3(x1) 20q11.22qter(x2) hmz Proportion of sub-clones with CNA/cnLOH <5% ~10% ~20% ~30% ~60% ~80% >90% Supplementary Figure 7. Schematic representation of the possible clonal architecture pre-treatment and at relapse for patient CLL092. Percentages below 10% and above 90% and cell populations without identifiable CNAs/cnLOHs are inferred. The precision of the percentage contribution of sub-clones with CNAs/cnLOH is +/- 10%. 32