Common variation at 3q26.2, 6p21.33, 17p11.2 and 22q13.1 influences multiple myeloma risk.

To identify variants for multiple myeloma risk, we conducted a genome-wide association study with validation in additional series totaling 4,692 individuals with multiple myeloma (cases) and 10,990 controls. We identified four risk loci at 3q26.2 (rs10936599, P = 8.70 × 10(-14)), 6p21.33 (rs2285803, PSORS1C2, P = 9.67 × 10(-11)), 17p11.2 (rs4273077, TNFRSF13B, P = 7.67 × 10(-9)) and 22q13.1 (rs877529, CBX7, P = 7.63 × 10(-16)). These data provide further evidence for genetic susceptibility to this B-cell hematological malignancy, as well as insight into the biological basis of predisposition.

Multiple myeloma (MM) is a malignancy of plasma cells1. Each year in the United States there are around 20,000 new cases of MM, and just over a half of that number die of the disease2. We have previously reported results of a genome-wide association study (GWAS) of MM based on an analysis of UK and German series and through fast track analysis of SNPs with the smallest P-values, identified risk loci at 2p23.3, 3p22.1 and 7p15.33. We have subsequently conducted further follow-up analyses, making use of an expanded German GWAS, and identified four new susceptibility loci for MM.

The German GWAS data set previously reported4, after QC comprised 1,014 MM cases recruited through Heidelberg University genotyped using Illumina OmniExpress BeadChips. Genotype frequencies were compared with genotype data generated by the Heinz-Nixdorf Recall (HNR) study of 2,107 individuals5 from the German population who had been genotyped using Illumina Human Omni1-Quad BeadChips and Illumina OmniExpress BeadChips (Online Methods).

The UK GWAS previously reported3, after QC comprised 1,321 MM cases recruited from the UK Medical Research Council (MRC) Myeloma-IX trial6 genotyped using Illumina OmniExpress BeadChips. Genotype frequencies were compared with publicly accessible genotype data generated by the UK Wellcome Trust Case Control Consortium 2 (WTCCC2) study of 2,698 individuals from the 1958 British Birth Cohort (known as 58C)7 and 2,501 individuals from the UK Blood Service (UKBS) collections that had been genotyped using Illumina Human 1.2M-Duo Custom_v1 Array BeadChips (Online Methods).

Genotype data from the GWAS were filtered on the basis of pre-specified quality-control measures (Online Methods). Individual SNPs were excluded from further analysis if they showed deviation from the Hardy–Weinberg equilibrium with a P<1.0x10-6 in controls, an individual SNP genotype yield <95%, or a minor allele frequency <1%. After filtering, 414,804 autosomal SNPs common to both case-control series were analyzable (Online Methods; Supplementary Figure 1; Supplementary Figure 2).

Prior to undertaking meta-analysis of the two GWAS, we searched for potential errors and biases in the datasets. Quantile-quantile plots of the genome-wide chi-squared values showed there was minimal inflation of the test statistics rendering substantial cryptic population substructure or differential genotype calling between cases and controls unlikely in either GWAS (genomic control inflation factor8, λgc=1.033 and 1.17 in UK and German GWAS, respectively; Supplementary Figure 3). For completeness principal components analysis was performed using the Eigenstrat9 software to determine the effects of population substructure on our findings (λcorrected=1.014 and 1.029 in UK and German GWAS, respectively; Supplementary Figure 3).

Using data on all cases and controls from both GWAS, we derived joint odds ratios (ORs) and confidence intervals (CIs) under a fixed effects model for each SNP and associated P-values10. In the combined analysis we identified nine SNPs showing good evidence of association (P<5.0×10−6) and mapping to distinct loci not previously associated with MM risk (Supplementary Table 1). The P-value threshold used does not exclude the possibility that other SNPs represent genuine association signals but was simply a pragmatic strategy for prioritizing replication.

To validate our findings, we conducted a replication study of the nine SNPs, genotyping samples from three additional series: UK-replication-1, 812 MM cases ascertained through the UK MRC Myeloma-IX and XI trials and 1,110 controls; UK-replication-2, 396 MM cases collected through UK haematology centers and 992 controls; German-replication 1,149 MM cases collected through the German Myeloma Study Group (DSMM), Heidelberg University Clinic and Ulm University Clinic, and 1,582 regional controls (Online Methods). In the combined analysis, four SNPs rs10936599 (3q26.2; P=8.70x10-14), rs2285803 (6p21.33; P= 9.67x10-11), rs4273077 (17p11.2; P=7.67x10-9) and rs877529 (22q13.1; P=7.63x10-16) showed evidence for an association with MM which was genome-wide significant (Table 1, Online Methods, Supplementary Table 2).

Table 1

Summary results for SNPs associated with multiple myeloma risk.

	Risk allele	RAFa	Case genotypes			RAFa	Control genotypes			ORb	95% CIc	P-value	Padjustedd
rs10936599 (3q26.2)	G		GG	AG	AA		GG	AG	AA
UK-GWAS		0.80	843	429	49	0.75	2960	1914	325	1.31	1.18-1.46	4.33x10-7	5.18x10-7
German-GWAS		0.79	632	332	49	0.75	1187	778	142	1.25	1.10-1.41	6.62x10-4	1.48x10-3
UK replication 1		0.80	520	259	30	0.76	628	415	63	1.32	1.13-1.55	4.98x10-4	-
UK replication 2		0.79	244	126	18	0.75	559	372	56	1.23	1.01-1.51	4.22x10-2	-
German replication		0.78	714	363	66	0.76	898	585	89	1.16	1.02-1.31	2.56x10-2	-
Combined										1.26	1.18-1.33	8.70x10-14	1.74x10-13
												Phet=0.60, I2=0%
rs2285803 (6p21.3)	A		AA	AG	GG		AA	AG	GG
UK-GWAS		0.32	125	603	593	0.28	444	2055	2699	1.21	1.10-1.32	6.67x10-5	7.64x10-5
German-GWAS		0.36	129	462	423	0.31	226	833	1047	1.24	1.11-1.39	1.15x10-4	1.18x10-4
UK replication 1		0.32	78	362	364	0.28	88	424	560	1.22	1.06-1.41	5.11x10-3	-
UK replication 2		0.29	32	152	193	0.26	51	402	510	1.14	0.94-1.38	1.82x10-1	-
German replication		0.33	130	491	521	0.30	140	674	752	1.12	1.00-1.26	5.86x10-2	-
Combined										1.19	1.13-1.26	9.67x10-11	1.18x10-10
												Phet=0.70, I2=0%
rs4273077 (17p11.2)	G		GG	AG	AA		GG	AG	AA
UK-GWAS		0.12	15	284	1022	0.10	48	926	4221	1.24	1.08-1.42	1.88x10-3	2.65x10-3
German-GWAS		0.14	25	239	750	0.11	27	390	1690	1.40	1.20-1.64	2.80x10-5	6.17x10-4
UK replication 1		0.12	18	148	629	0.09	12	179	915	1.28	1.04-1.57	1.96x10-2	-
UK replication 2		0.11	3	77	304	0.10	8	178	805	1.12	0.85-1.48	4.20x10-1	-
German replication		0.12	17	252	876	0.11	21	298	1244	1.17	0.99-1.38	6.79x10-2	-
Combined										1.26	1.16-1.36	7.67x10-9	1.41x10-7
												Phet=0.50 I2=0%
rs877529 (22q13.1)	A		AA	AG	GG		AA	AG	GG
UK-GWAS		0.51	346	654	321	0.44	1000	2560	1633	1.33	1.22-1.45	1.08x10-10	9.11x10-11
German-GWAS		0.45	214	483	317	0.43	389	1026	692	1.09	0.98-1.21	1.18x10-1	1.09x10-1
UK replication 1		0.49	176	436	192	0.44	195	555	322	1.24	1.08-1.41	2.01x10-3	-
UK replication 2		0.47	86	192	109	0.42	166	485	327	1.24	1.05-1.47	1.21x10-2	-
German replication		0.46	238	586	321	0.41	274	754	544	1.23	1.10-1.37	2.69x10-4	-
Combined										1.23	1.17-1.29	7.63x10-16	2.29x10-16
												Phet=0.09, I2=51%

Risk allele frequency (RAF).

Odds ratio.

95% Confidence Interval.

Eigenstrat adjusted P-values.

rs10936599 at 3q26.2 (P=8.70x10-14; Table 1) is responsible for the H717H polymorphism in the myoneurin gene (MYNN; MIM 606042). The rs10936599 G risk allele has previously also been shown to influence colorectal cancer risk11. While MYNN encodes a zinc finger protein of unknown function expressed principally in muscle rs10936599 (169,492,101bps) however maps within a 250Kb region of LD which also encompasses the telomerase RNA component gene (TERC; MIM 602322). Telomerase reactivation and telomerase-mediated elongation of shorter telomeres is a feature of MM12. Since carrier status for the rs10936599 G risk allele is associated with significantly longer telomeres12 TERC represents an attractive candidate for MM susceptibility. Moreover imputation of untyped genotypes in cases and controls using 1000 genomes data provided for a marginally stronger association at 3q26.2 with A allele of rs2293607, which maps 63bps 5’ to TERC (P=6.2x10-10 compared with 1.3x10-9; for rs10936599 in meta-analysis of GWAS data; Figure 1). The A allele of rs2293607 has recently been shown to be associated with TERC mRNA expression and longer telomeres in vitro12 supporting variation in TERC as the basis of the 3q26.2 cancer association.

Figure 1

Regional plots of association results and recombination rates for the 3q26.2, 6p21.33, 17p11.2 and 22q13.1 susceptibility loci.

(a-d) Association results of both genotyped (triangles) and imputed (circles) SNPs in the GWAS samples and recombination rates for rates within the four loci: 3q26.2, 6p21.33, 17p11.2 and 22q13.1. For each plot, −log10 P values (y axis) of the SNPs are shown according to their chromosomal positions (x axis). The top genotyped SNP in each combined analysis is a large triangle and is labeled by its rsID. The color intensity of each symbol reflects the extent of LD with the top genotyped SNP: white (r2=0) through to dark red (r2=1.0). Genetic recombination rates (cM/Mb), estimated using HapMap CEU samples, are shown with a light blue line. Physical positions are based on NCBI build 36 of the human genome. Also shown are the relative positions of genes and transcripts mapping to each region of association. Genes have been redrawn to show the relative positions; therefore, maps are not to physical scale. The lower panel shows the region of interest together with all transcripts and chromatin state segmentation track (ChromHMM) for lymphoblastoid cells using data from the HapMap Encode Project.

rs2285803 (P=9.67x10-11; Table 1; Figure 1) localizes in intron 5 of the putative psoriasis susceptibility gene, PSORS1C1 (MIM 613525) at 6p21.33 (31,107,245bps). The 163 kb region of LD also encompasses CCHCR1 (MIM 605310), CDSN (MIM 602593), transcription factor 19 (TCF19; MIM 600912) and POU domain, class 5, transcription factor 1 (POU5F1, MIM 164177) genes. While there is currently no evidence for POU5F1 playing a role in MM intriguingly the gene encodes OCT3/OCT4 which regulates pluripotency, lineage commitment and regulates tissue-specific gene expression. Variation at 6p21.33 has previously been shown to be associated with follicular lymphoma (FL) and Hodgkin lymphoma (HL) risk. The associations for FL defined by rs6457327 in the HLA class I region13 and rs10484561 and rs2647012 in the HLA class II region14–15. The HL association at 6p21.33 is marked by rs6903608 in the HLA class II region16. The risk of MM associated with each of these SNPs was non-significant (Supplementary Table 3). To further investigate the rs2285803 signal for MM we imputed classical HLA alleles from SNP data from both GWASs using HLA*IMP17–18. The strongest HLA association was provided by HLA-DRB5*01 (P=1.42x10-5; Supplementary Table 4) which was significantly weaker than provided by rs2285803 (P=3.07x10-8). To evaluate the independence of associations, we conducted regression, jointly on rs2285803 and the imputed HLA alleles. Conditional analysis showed that most, but not all, of the MHC variation defined by SNP genotype could be explained for by rs2285803 (Supplementary Table 4).

rs4273077 (P=7.67x10-9; Table 1) maps within intron 2 of the gene for Homo sapiens tumor necrosis factor receptor superfamily member 13B (TNFRSF13B; MIM 604907 at 17p11.2 (16,849,139bps; Figure 1). TNFRSF13B (alias TACI), represents a strong candidate for MM predisposition a priori. TNFRSF13B is a key regulator of B and T-cell function being required for the development of transitional (T2) and mature-B lymphocytes, and regulation of normal B-cell homeostasis19. Variation at TNFRSF13B influences circulating IgG levels20 and Tnfrsf13b-/- mice show an expanded B-cell population with lymphoproliferation and lymphoma risk21. Since TNFRSF13B mutation is a risk factor for antibody-deficient (MIM 240500) and selective Ig deficiency (MIM 609529) associated with lymphoproliferation it is likely that loss of TNFRSF13B function impairs isotype switching. Primary MM cells with a high TNFRSF13B expression (TACIhigh) resemble bone marrow plasma cells which depend on the interaction with the bone marrow environment. In contrast MM cells with a low expression of TNFRSF13B (TACIlow) resemble plasmablasts22. TACI-Ig, a soluble receptor blocking the TNFRSF13B ligands BAF and APRIL, inhibits the growth of TACIhigh but not TACIlow myeloma cells in the SCID-hu model23.

rs877529 localizes to intron 2 of the gene encoding chromobox homolog 7 (CBX7; MIM 608457) at 22q13.1 (39,542,292bps; P=7.63x10-16; Table 1; Figure 1). CBX7 encodes a polycomb group protein. These proteins form part of a gene regulatory mechanism that determines cell fate during development as well as contributing to the control of normal cell growth and differentiation24. CBX7-mediated repression of transcription acts through Ink4a/Arf25, cooperating with Myc to promote aggressive B-cell lymphomagenesis with high levels of CBX7 being a feature of germinal center-derived follicular lymphoma26.

To explore whether any of the associations reflect cis-acting regulatory effects we studied mRNA expression in CD138-selected plasma cells27 and lymphoblastoid cell lines (LCLs)28–30 (Online Methods; Supplementary Table 5). Although we found no association between genotype and expression of either mRNA transcript, steady-state levels of RNA at a single time point may not adequately capture the impact of differential expression in tumourigenesis. To explore epigenetic profile of association signals we made use of chromatin state segmentation in lymphoblastoid cell lines data generated by the Encode Project31. rs2293607 maps to a region of active chromatin predicted by ENCODE data to be an active promoter and rs877529 maps within a strong enhancer element within CBX7 (Figure 1).

Hierarchically MM can be broadly divided into hyperdiploid and non-hyperdiploid subtypes32–33. The latter is primarily composed of patients harboring IGH translocations, principally t(11;14)(q13;q32) and t(4;14)(p16;q32)34,35. Case-only analysis provided no evidence for a subtype specific association with genotype for rs10936599, rs2285803 or rs4237077 consistent with each variant having a generic effect on MM risk (Supplementary Table 6). In contrast rs877529 showed evidence, significant after correction for multiple testing, that the association is driven by non t(11;14) MM (P=8.0x10-4; Padj= 0.016).

Our findings provide further evidence for inherited genetic susceptibility to MM and insight into the development of this hematological malignancy. We estimate that the seven loci we have so far identified account for ~13% of the familial risk of MM. While the power of our study to detect the major common loci conferring risks of ≥1.3 was high we had low power to detect alleles with smaller effects and/or minor allele frequencies (MAFs) <0.1. By implication, variants with such profiles are likely to represent a much larger class of susceptibility loci for MM, because of truly small effect sizes or submaximal LD with tagging SNPs. Thus, it is likely that a large number of variants remain to be discovered. This assertion is supported by the continued excess of associations observed over those expected, in addition to the regions studied herein. Further efforts to expand the scale of GWAS, in terms of both sample size and SNP coverage, and to increase the number of SNPs taken forward to large-scale replication may therefore identify additional risk variants. Finally as we have recently shown stratified analysis of MM by karyotype may lead additional subtype-specific risk variants4.

Online Methods

Ethics

Collection of samples and clinico-pathological information from subjects was undertaken with informed consent and relevant ethical review board approval in accordance with the tenets of the Declaration of Helsinki.

Genome-wide association study

UK-GWAS: Details of this study have been previously reported3. Briefly, 1,371 MM (ICD-10 C90.0; 469 male; mean age at diagnosis 63.9 years, SD 9.9) were ascertained through the UK Medical Research Council (MRC) Myeloma-IX trial6. Genotyping of cases was performed using Illumina Human OmniExpress-12 v1.0 arrays according to the manufacturer's protocols (Illumina, San Diego, USA). For controls, we used publicly accessible data generated by the Wellcome Trust Case Control Consortium from the 1958 Birth Cohort (58C; also known as the National Child Development Study)7 and National Blood Service (NBS). Genotyping of controls was conducted using Illumina Human 1-2M-Duo Custon_v1 Array chips. SNP calling was performed using Illuminus Software. Full details of genotyping, SNP calling and QC have been previously reported (www.wtccc.org.uk).

German-GWAS: The German-GWAS comprised 384 MM cases (229 male, mean age at diagnosis 54.5 years, SD 8.0) which were the subject of a previous publication3 and an additional series of 698 MM cases (389 male, mean age at diagnosis of 59 years; SD 9.3) recruited by the German Multiple Myeloma Study Group (GMMG), coordinated by the University Clinic, Heidelberg. All cases were genotyped using Illumina Human OmniExpress-12 v1.0 arrays according to the manufacturer's protocols (Illumina, San Diego, USA). For controls, we used genotype data on 2,132 healthy individuals, enrolled into the Heinz Nixdorf Recall (HNR) study5; of these 704 were genotyped using Illumina HumanOmni1-Quad_v1 and 1428 OmniExpress-12 v1.0.

Quality control of GWAS datasets

DNA samples with GenCall scores <0.25 at any locus were considered “no calls”. A SNP was deemed to have failed if <95% of DNA samples generated a genotype at the locus. Cluster plots were manually inspected for all SNPs considered for replication. The same quality control metrics on the new German GWAS data were applied as in our previous MM study3. We restricted analyses to samples for whom >95% of SNPs were successfully genotyped, thus eliminating 10 samples (Supplementary Figure 1). We computed identity-by-state (IBS) probabilities for all pairs (cases and controls) to search for duplicates and closely related individuals amongst samples (defined as IBS ≥0.80, thereby excluding first-degree relatives). For all identical pairs the sample having the highest call rate was retained, eliminating 13 samples. To identify individuals who might have non-Western European ancestry, we merged our case and control data with phase II HapMap samples (60 western European [CEU], 60 Nigerian [YRI], 90 Japanese [JPT] and 90 Han Chinese [CHB]). For each pair of individuals we calculated genome-wide IBS distances on markers shared between HapMap and our SNP panel, and used these as dissimilarity measures upon which to perform principal component analysis. The first two principal components for each individual were plotted and any individual not present in the main CEU cluster was excluded from analyses. We removed 70 samples of non-CEU ancestry (some of which had poor call rates). We filtered out SNPs having a minor allele frequency [MAF] <1%, and a call rate <95% in cases or controls. We also excluded SNPs showing departure from Hardy-Weinberg equilibrium (HWE) at P<10-6 in controls. For replication and validation analysis call rates were >95% per 384-well plate for each SNP; cluster plots were visually examined by two researchers.

Replication series and genotyping

UK-replication-1 comprised 812 MM cases (412 male) collected through the UK Medical Research Council (MRC) Myeloma-IX (n=95) and XI trials (n=717). Controls comprised 1,110 healthy individuals with self reported European ancestry (420 male, aged 18-69 years) with no personal history of malignancy ascertained through GEnetic Lung CAncer Predisposition Study (GELCAPS; n=536)36 and National Study of Colorectal Cancer Genetics (NSCCG; n=574)37 studies. All cases and controls were UK subjects.

UK-replication-2 comprised 396 MM cases (181 male; mean age at diagnosis 66.0 years, SD 12.5) collected through UK haematology departments (2001-present) including the Royal Marsden Hospitals NHS Trust (RMH). Controls were 992 healthy individuals (421 male, mean age 57.4 years, SD 12.3) with no personal history of malignancy who were the spouses of cancer patients ascertained by the ICR between 2000 and 2008.

German-replication comprised 1,149 cases collected by the German Myeloma Study Group (DSMM), GMMG, University Clinic, Heidelberg and University Clinic, Ulm (676 males; mean age at diagnosis 57.6 years, SD 9.8). The control population was composed of 1,582 healthy German blood donors who were recruited between 2004 and 2007 by the Institute of Transfusion Medicine and Immunology, University of Mannheim, Germany (885 male, mean age 55.8 years, SD 10.0).

Replication genotyping was performed using competitive allele-specific PCR KASPar chemistry (KBiosciences Ltd, Hertfordshire, UK). All primers and probes used are available on request. Samples having SNP call rates of <90% were excluded from the analysis. To ensure quality of genotyping in all assays, at least two negative controls and 1-2% duplicates (showing a concordance >99.99%) were genotyped. To exclude technical artifact in genotyping we performed cross-platform validation of 384 samples and sequenced a set of 384 randomly selected samples from each case and control series to confirm genotyping accuracy (concordance >99.9%).

Sample preparation

For German cases DNA was prepared from EDTA-venous blood samples, 100% of the original GWAS, 42% of the additional GWAS, and 91% of the replication samples; for the remaining cases, the source was the CD138-negative fraction of bone marrow cells, with < 5% contamination by tumor cells. For all UK cases DNA was prepared from EDTA-venous blood samples. Samples were obtained prior to delivery of chemotherapy in the vast majority of UK cases and at least 80% of the German cases. All DNAs were extracted using Qiagen FlexiGene or QIAamp methodologies and quantified using PicoGreen (Invitrogen).

Statistical and bioinformatic analysis

Main analyses were undertaken using R (v2.6), Stata v.10 (State College, Texas, US) and PLINK (v1.06)38 software. Odds ratios (ORs) and associated 95% confidence intervals (CIs) along with associated P-values were calculated by unconditional logistic regression. The adequacy of the case-control matching and possibility of differential genotyping of cases and controls were formally evaluated using quantile-quantile (Q-Q) plots of test statistics. The inflation factor λ was based on the 90% least significant SNPs8. We undertook adjustment for possible population substructure using Eigenstrat software. Meta-analysis was conducted using standard methods10. Cochran’s Q statistic to test for heterogeneity10 and the I statistic to quantify the proportion of the total variation due to heterogeneity were calculated39. I values ≥75% are considered characteristic of large heterogeneity39–40. To conduct a pooled analysis incorporating Eigenstrat adjusted P-values from the GWAS we used the weighted Z-method implemented in the program METAL41. Since not all the HNR controls were genotyped using the same Illumina array the robustness of genomewide associations was formally assessed by deriving ORs for the different German case-control combinations and incorporation of these data in meta-analysis (Supplementary Figure 4). We examined each SNP for dose response by comparing 1-d.f. and 2-d.f. logistic regression models, adjusting for stage using a likelihood ratio test, and examined the combined effects of multiple SNPs by evaluating the effect of adding an interaction term on the model by using a likelihood ratio test and adjusting for stage. Associations by tumor karyotype were examined by logistic regression in case-only analyses. The sibling relative risk attributable to a given SNP was calculated using the formula: where p is the population frequency of the minor allele, q=1−p, and r1 and r2 are the relative risks (estimated as OR) for heterozygotes and rare homozygotes, relative to common homozygotes. Assuming a multiplicative interaction, the proportion of the familial risk attributable to a SNP was calculated as logλ*/logλ0, where λ0 is the overall familial relative risk estimated from epidemiological studies, assumed to be 2.4542.

Prediction of the untyped SNPs was carried out using IMPUTEv2, based on the 1000 genomes phase 1 integrated variant set (b37) from March 2012. Imputed data were analysed using SNPTEST v2 to account for uncertainties in SNP prediction and meta-analysis was performed using METAv1.443. LD metrics were calculated in plink using 1000 genomes data and plotted using SNAP. LD blocks were defined on the basis of HapMap recombination rate (cM/Mb) as defined using the Oxford recombination hotspots44 and on the basis of distribution of confidence intervals defined by Gabriel et al45. We imputed classical HLA alleles from GWAS SNPs using HLA*IMP17–18

To explore epigenetic profile of association signals we made use of chromatin state segmentation in lymphoblastoid cell lines data generated by the Encode Project31. The states were inferred from ENCODE Histone Modification data (H4K20me1, H3K9ac, H3K4me3, H3K4me2, H3K4me1, H3K36me3, H3K27me3, H3K27ac and CTCF) binarized using a multivariate Hidden Markov Model.

Karotyping and Fluorescence in situ hybridization (FISH)

Conventional cytogenetic studies of multiple myeloma cells were conducted using standard karotyping methodologies, and standard criteria for the definition of a clone were applied. FISH and ploidy classification of UK samples was conducted using the methodology described by Chiecchio et al . FISH and ploidy classification of German samples was performed as previously described47. The XL IGH Break Apart probe (MetaSystems, Altlussheim Germany) was used to detect any IGH translocation in German samples.

Relationship between SNP genotype and mRNA expression

To examine for a relationship between SNP genotype and mRNA expression in MM we made use of Affymetrix Human Genome U133+2.0 array data on the plasma cells from 192 MM patients from the MRC Myeloma IX trial27. To assay TERC which was not captured on the U133+2.0 array we made use of Affymetrix GeneChip miRNA 2.0 data. To examine for a relationship between SNP genotype and expression levels in lymphocytes we made use of publicly available expression data generated on lymphoblastoid cell lines from HapMap329 Geneva GenCord individuals30 and the MuTHER resource28 using Sentrix Human-6 Expression BeadChips (Illumina, San Diego, USA)48–49.

Supplementary Material