Unsupervised detection of fragment length signatures of circulating tumor DNA using non-negative matrix factorization.

Sequencing of cell-free DNA (cfDNA) is currently being used to detect cancer by searching both for mutational and non-mutational alterations. Recent work has shown that the length distribution of cfDNA fragments from a cancer patient can inform tumor load and type. Here, we propose non-negative matrix factorization (NMF) of fragment length distributions as a novel and completely unsupervised method for studying fragment length patterns in cfDNA. Using shallow whole-genome sequencing (sWGS) of cfDNA from a cohort of patients with metastatic castration-resistant prostate cancer (mCRPC), we demonstrate how NMF accurately infers the true tumor fragment length distribution as an NMF component - and that the sample weights of this component correlate with ctDNA levels (r=0.75). We further demonstrate how using several NMF components enables accurate cancer detection on data from various early stage cancers (AUC = 0.96). Finally, we show that NMF, when applied across genomic regions, can be used to discover fragment length signatures associated with open chromatin.

Expression patterns and prognostic potential of circular RNAs in mantle cell lymphoma: a study of younger patients from the MCL2 and MCL3 clinical trials.

Mantle cell lymphoma (MCL) is characterized by marked differences in outcome, emphasizing the need for strong prognostic biomarkers. Here, we explore expression patterns and prognostic relevance of circular RNAs (circRNAs), a group of endogenous non-coding RNA molecules, in MCL. We profiled the circRNA expression landscape using RNA-sequencing and explored the prognostic potential of 40 abundant circRNAs in samples from the Nordic MCL2 and MCL3 clinical trials, using NanoString nCounter Technology. We report a circRNA-based signature (circSCORE) developed in the training cohort MCL2 that is highly predictive of time to progression (TTP) and lymphoma-specific survival (LSS). The dismal outcome observed in the large proportion of patients assigned to the circSCORE high-risk group was confirmed in the independent validation cohort MCL3, both in terms of TTP (HR 3.0; P = 0.0004) and LSS (HR 3.6; P = 0.001). In Cox multiple regression analysis incorporating MIPI, Ki67 index, blastoid morphology and presence of TP53 mutations, circSCORE retained prognostic significance for TTP (HR 3.2; P = 0.01) and LSS (HR 4.6; P = 0.01). In conclusion, circRNAs are promising prognostic biomarkers in MCL and circSCORE improves identification of high-risk disease among younger patients treated with cytarabine-containing chemoimmunotherapy and autologous stem cell transplant.

Hierarchical Classification of Cancers of Unknown Primary Using Multi-Omics Data.

A cancer of unknown primary (CUP) is a metastatic cancer for which standard diagnostic tests fail to locate the primary cancer. As standard treatments are based on the cancer type, such cases are hard to treat and have very poor prognosis. Using molecular data from the metastatic cancer to predict the primary site can make treatment choice easier and enable targeted therapy. In this article, we first examine the ability to predict cancer type using different types of omics data. Methylation data lead to slightly better prediction than gene expression and both these are superior to classification using somatic mutations. After using 3 data types independently, we notice some differences between the classes that tend to be misclassified, suggesting that integrating the data might improve accuracy. In light of the different levels of information provided by different omics types and to be able to handle missing data, we perform multi-omics classification by hierarchically combining the classifiers. The proposed hierarchical method first classifies based on the most informative type of omics data and then uses the other types of omics data to classify samples that did not get a high confidence classification in the first step. The resulting hierarchical classifier has higher accuracy than any of the single omics classifiers and thus proves that the combination of different data types is beneficial. Our results show that using multi-omics data can improve the classification of cancer types. We confirm this by testing our method on metastatic cancers from the MET500 dataset.

A novel combined miRNA and methylation marker panel (miMe) for prediction of prostate cancer outcome after radical prostatectomy.

Improved prognostic biomarkers are needed to guide personalized prostate cancer (PC) treatment decisions. Due to the prominent molecular heterogeneity of PC, multimarker panels may be more robust. Here, 25 selected top-candidate miRNA and methylation markers for PC were profiled by qPCR in malignant radical prostatectomy (RP) tissue specimens from 198 PC patients (Cohort 1, training). Using GLMnet, we trained a novel multimarker model (miMe) comprising nine miRNAs and three methylation markers that predicted postoperative biochemical recurrence (BCR) independently of the established clinicopathological CAPRA-S nomogram in Cox multivariate regression analysis in Cohort 1 (HR [95% CI]: 1.53 [1.26-1.84], p < 0.001). This result was successfully validated in two independent RP cohorts (Cohort 2, n = 159: HR [95% CI]: 1.35 [1.06-1.73], p = 0.015. TCGA, n = 350: HR [95% CI]: 1.34 [1.01-1.77], p = 0.04). Notably, in CAPRA-S low-risk patients, a high miMe score was associated with >6 times higher risk of BCR, suggesting that miMe may help identify PC patients at high risk of progression despite favorable clinicopathological factors postsurgery. Finally, miMe was a significant predictor of cancer-specific survival (p = 0.019, log-rank test) in a merged analysis of 357 RP patients. In conclusion, we trained, tested and validated a novel 12-marker panel (miMe) that showed significant independent prognostic value in three RP cohorts. In the future, combining miMe score with existing clinical nomograms may improve PC risk stratification and thus help guide treatment decisions.

Proteomic profiling identifies outcome-predictive markers in patients with peripheral T-cell lymphoma, not otherwise specified.

Peripheral T-cell lymphoma, not otherwise specified (PTCL-NOS) constitutes a heterogeneous category of lymphomas, which do not fit into any of the specifically defined T-cell lymphoma entities. Both the pathogenesis and tumor biology in PTCL-NOS are poorly understood. Protein expression in pretherapeutic PTCL-NOS tumors was analyzed by proteomics. Differentially expressed proteins were compared in 3 distinct scenarios: (A) PTCL-NOS tumor tissue (n = 18) vs benign lymphoid tissue (n = 8), (B) clusters defined by principal component analysis (PCA), and (C) tumors from patients with chemosensitive vs refractory PTCL-NOS. Selected differentially expressed proteins identified by proteomics were correlated with clinico-pathological features and outcome in a larger cohort of patients with PTCL-NOS (n = 87) by immunohistochemistry (IHC). Most proteins with altered expression were identified comparing PTCL-NOS vs benign lymphoid tissue. PCA of the protein profile defined 3 distinct clusters. All benign samples clustered together, whereas PTCL-NOS tumors separated into 2 clusters with different patient overall survival rates (P = .001). Differentially expressed proteins reflected large biological diversity among PTCL-NOS, particularly associated with alterations of "immunological" pathways. The 2 PTCL-NOS subclusters defined by PCA showed disturbance of "stress-related" and "protein metabolic" pathways. α-Enolase 1 (ENO1) was found differentially expressed in all 3 analyses, and high intratumoral ENO1 expression evaluated by IHC correlated with poor outcome (hazard ratio, 2.09; 95% confidence interval, 1.17-3.73; P = .013). High expression of triosephosphate isomerase (TPI1) also showed a tendency to correlate with poor survival (P = .057). In conclusion, proteomic profiling of PTCL-NOS provided evidence of markedly altered protein expression and identified ENO1 as a novel potential prognostic marker.

A site specific model and analysis of the neutral somatic mutation rate in whole-genome cancer data.

BACKGROUND: Detailed modelling of the neutral mutational process in cancer cells is crucial for identifying driver mutations and understanding the mutational mechanisms that act during cancer development. The neutral mutational process is very complex: whole-genome analyses have revealed that the mutation rate differs between cancer types, between patients and along the genome depending on the genetic and epigenetic context. Therefore, methods that predict the number of different types of mutations in regions or specific genomic elements must consider local genomic explanatory variables. A major drawback of most methods is the need to average the explanatory variables across the entire region or genomic element. This procedure is particularly problematic if the explanatory variable varies dramatically in the element under consideration. RESULTS: To take into account the fine scale of the explanatory variables, we model the probabilities of different types of mutations for each position in the genome by multinomial logistic regression. We analyse 505 cancer genomes from 14 different cancer types and compare the performance in predicting mutation rate for both regional based models and site-specific models. We show that for 1000 randomly selected genomic positions, the site-specific model predicts the mutation rate much better than regional based models. We use a forward selection procedure to identify the most important explanatory variables. The procedure identifies site-specific conservation (phyloP), replication timing, and expression level as the best predictors for the mutation rate. Finally, our model confirms and quantifies certain well-known mutational signatures. CONCLUSION: We find that our site-specific multinomial regression model outperforms the regional based models. The possibility of including genomic variables on different scales and patient specific variables makes it a versatile framework for studying different mutational mechanisms. Our model can serve as the neutral null model for the mutational process; regions that deviate from the null model are candidates for elements that drive cancer development.

Prognostic miRNA classifier in early-stage mycosis fungoides: development and validation in a Danish nationwide study.

Mycosis fungoides (MF) is the most frequent form of cutaneous T-cell lymphoma. The disease often takes an indolent course, but in approximately one-third of the patients, the disease progresses to an aggressive malignancy with a poor prognosis. At the time of diagnosis, it is impossible to predict which patients develop severe disease and are in need of aggressive treatment. Accordingly, we investigated the prognostic potential of microRNAs (miRNAs) at the time of diagnosis in MF. Using a quantitative reverse transcription polymerase chain reaction platform, we analyzed miRNA expression in diagnostic skin biopsies from 154 Danish patients with early-stage MF. The patients were subdivided into a discovery cohort (n = 82) and an independent validation cohort (n = 72). The miRNA classifier was built using a LASSO (least absolute shrinkage and selection operator) Cox regression to predict progression-free survival (PFS). We developed a 3-miRNA classifier, based on miR-106b-5p, miR-148a-3p, and miR-338-3p, which successfully separated patients into high-risk and low-risk groups of disease progression. PFS was significantly different between these groups in both the discovery cohort and the validation cohort. The classifier was stronger than existing clinical prognostic factors and remained a strong independent prognostic tool after stratification and adjustment for these factors. Importantly, patients in the high-risk group had a significantly reduced overall survival. The 3-miRNA classifier is an effective tool to predict disease progression of early-stage MF at the time of diagnosis. The classifier adds significant prognostic value to existing clinical prognostic factors and may facilitate more individualized treatment of these patients.

Prediction of Primary Tumors in Cancers of Unknown Primary.

A cancer of unknown primary (CUP) is a metastatic cancer for which standard diagnostic tests fail to identify the location of the primary tumor. CUPs account for 3-5% of cancer cases. Using molecular data to determine the location of the primary tumor in such cases can help doctors make the right treatment choice and thus improve the clinical outcome. In this paper, we present a new method for predicting the location of the primary tumor using gene expression data: locating cancers of unknown primary (LoCUP). The method models the data as a mixture of normal and tumor cells and thus allows correct classification even in impure samples, where the tumor biopsy is contaminated by a large fraction of normal cells. We find that our method provides a significant increase in classification accuracy (95.8% over 90.8%) on simulated low-purity metastatic samples and shows potential on a small dataset of real metastasis samples with known origin.

Pathway Analysis of Skin from Psoriasis Patients after Adalimumab Treatment Reveals New Early Events in the Anti-Inflammatory Mechanism of Anti-TNF-α.

Psoriasis is a chronic cutaneous inflammatory disease. The immunopathogenesis is a complex interplay between T cells, dendritic cells and the epidermis in which T cells and dendritic cells maintain skin inflammation. Anti-tumour necrosis factor (anti-TNF)-α agents have been approved for therapeutic use across a range of inflammatory disorders including psoriasis, but the anti-inflammatory mechanisms of anti-TNF-α in lesional psoriatic skin are not fully understood. We investigated early events in skin from psoriasis patients after treatment with anti-TNF-α antibodies by use of bioinformatics tools. We used the Human Gene 1.0 ST Array to analyse gene expression in punch biopsies taken from psoriatic patients before and also 4 and 14 days after initiation of treatment with the anti-TNF-α agent adalimumab. The gene expression was analysed by gene set enrichment analysis using the Functional Annotation Tool from DAVID Bioinformatics Resources. The most enriched pathway was visualised by the Pathview Package on Kyoto Encyclopedia of Genes and Genomes (KEGG) graphs. The analysis revealed new very early events in psoriasis after adalimumab treatment. Some of these events have been described after longer periods of anti-TNF-α treatment when clinical and histological changes appear, suggesting that effects of anti-TNF-α treatment on gene expression appear very early before clinical and histological changes. Combining microarray data on biopsies from psoriasis patients with pathway analysis allowed us to integrate in vitro findings into the identification of mechanisms that may be important in vivo. Furthermore, these results may reflect primary effect of anti-TNF-α treatment in contrast to studies of gene expression changes following clinical and histological changes, which may reflect secondary changes correlated to the healing of the skin.

Identifying disease associated genes by network propagation.

BACKGROUND: Genome-wide association studies have identified many individual genes associated with complex traits. However, pathway and network information have not been fully exploited in searches for genetic determinants, and including this information may increase our understanding of the underlying biology of common diseases. RESULTS: In this study, we propose a framework to address this problem in a principled way, with the underlying hypothesis that complex disease operates through multiple connected genes. Associations inferred from GWAS are translated into prior scores for vertices in a protein-protein interaction network, and these scores are propagated through the network. Permutation is used to select genes that are guilty-by-association and thus consistently obtain high scores after network propagation. We apply the approach to data of Crohn's disease and call candidate genes that have been reported by other independent GWAS, but not in the analysed data set. A prediction model based on these candidate genes show good predictive power as measured by Area Under the Receiver Operating Curve (AUC) in 10 fold cross-validations. CONCLUSIONS: Our network propagation method applied to a genome-wide association study increases association findings over other approaches.

Mutations in BRIP1 confer high risk of ovarian cancer.

Ovarian cancer causes more deaths than any other gynecologic malignancy in developed countries. Sixteen million sequence variants, identified through whole-genome sequencing of 457 Icelanders, were imputed to 41,675 Icelanders genotyped using SNP chips, as well as to their relatives. Sequence variants were tested for association with ovarian cancer (N of affected individuals = 656). We discovered a rare (0.41% allelic frequency) frameshift mutation, c.2040_2041insTT, in the BRIP1 (FANCJ) gene that confers an increase in ovarian cancer risk (odds ratio (OR) = 8.13, P = 2.8 × 10(-14)). The mutation was also associated with increased risk of cancer in general and reduced lifespan by 3.6 years. In a Spanish population, another frameshift mutation in BRIP1, c.1702_1703del, was seen in 2 out of 144 subjects with ovarian cancer and 1 out of 1,780 control subjects (P = 0.016). This allele was also associated with breast cancer (seen in 6/927 cases; P = 0.0079). Ovarian tumors from heterozygous carriers of the Icelandic mutation show loss of the wild-type allele, indicating that BRIP1 behaves like a classical tumor suppressor gene in ovarian cancer.

European genome-wide association study identifies SLC14A1 as a new urinary bladder cancer susceptibility gene.

Three genome-wide association studies in Europe and the USA have reported eight urinary bladder cancer (UBC) susceptibility loci. Using extended case and control series and 1000 Genomes imputations of 5 340 737 single-nucleotide polymorphisms (SNPs), we searched for additional loci in the European GWAS. The discovery sample set consisted of 1631 cases and 3822 controls from the Netherlands and 603 cases and 37 781 controls from Iceland. For follow-up, we used 3790 cases and 7507 controls from 13 sample sets of European and Iranian ancestry. Based on the discovery analysis, we followed up signals in the urea transporter (UT) gene SLC14A. The strongest signal at this locus was represented by a SNP in intron 3, rs17674580, that reached genome-wide significance in the overall analysis of the discovery and follow-up groups: odds ratio = 1.17, P = 7.6 × 10(-11). SLC14A1 codes for UTs that define the Kidd blood group and are crucial for the maintenance of a constant urea concentration gradient in the renal medulla and, through this, the kidney's ability to concentrate urine. It is speculated that rs17674580, or other sequence variants in LD with it, indirectly modifies UBC risk by affecting urine production. If confirmed, this would support the 'urogenous contact hypothesis' that urine production and voiding frequency modify the risk of UBC.

Genome-wide significant association between a sequence variant at 15q15.2 and lung cancer risk.

Genome-wide association studies (GWAS) have identified 3 genomic regions, at 15q24-25.1, 5p15.33, and 6p21.33, which associate with the risk of lung cancer. Large meta-analyses of GWA data have failed to find additional associations of genome-wide significance. In this study, we sought to confirm 7 variants with suggestive association to lung cancer (P < 10(-5)) in a recently published meta-analysis. In a GWA dataset of 1,447 lung cancer cases and 36,256 controls in Iceland, 3 correlated variants on 15q15.2 (rs504417, rs11853991, and rs748404) showed a significant association with lung cancer, whereas rs4254535 on 2p14, rs1530057 on 3p24.1, rs6438347 on 3q13.31, and rs1926203 on 10q23.31 did not. The most significant variant, rs748404, was genotyped in an additional 1,299 lung cancer cases and 4,102 controls from the Netherlands, Spain, and the United States and the results combined with published GWAS data. In this analysis, the T allele of rs748404 reached genome-wide significance (OR = 1.15, P = 1.1 × 10(-9)). Another variant at the same locus, rs12050604, showed association with lung cancer (OR = 1.09, 3.6 × 10(-6)) and remained significant after adjustment for rs748404 and vice versa. rs748404 is located 140 kb centromeric of the TP53BP1 gene that has been implicated in lung cancer risk. Two fully correlated, nonsynonymous coding variants in TP53BP1, rs2602141 (Q1136K) and rs560191 (E353D) showed association with lung cancer in our sample set; however, this association did not remain significant after adjustment for rs748404. Our data show that 1 or more lung cancer risk variants of genome-wide significance and distinct from the coding variants in TP53BP1 are located at 15q15.2.

Genetic correction of PSA values using sequence variants associated with PSA levels.

Measuring serum levels of the prostate-specific antigen (PSA) is the most common screening method for prostate cancer. However, PSA levels are affected by a number of factors apart from neoplasia. Notably, around 40% of the variability of PSA levels in the general population is accounted for by inherited factors, suggesting that it may be possible to improve both sensitivity and specificity by adjusting test results for genetic effects. To search for sequence variants that associate with PSA levels, we performed a genome-wide association study and follow-up analysis using PSA information from 15,757 Icelandic and 454 British men not diagnosed with prostate cancer. Overall, we detected a genome-wide significant association between PSA levels and single-nucleotide polymorphisms (SNPs) at six loci: 5p15.33 (rs2736098), 10q11 (rs10993994), 10q26 (rs10788160), 12q24 (rs11067228), 17q12 (rs4430796), and 19q13.33 [rs17632542 (KLK3: I179T)], each with P(combined) <3 × 10(-10). Among 3834 men who underwent a biopsy of the prostate, the 10q26, 12q24, and 19q13.33 alleles that associate with high PSA levels are associated with higher probability of a negative biopsy (odds ratio between 1.15 and 1.27). Assessment of association between the six loci and prostate cancer risk in 5325 cases and 41,417 controls from Iceland, the Netherlands, Spain, Romania, and the United States showed that the SNPs at 10q26 and 12q24 were exclusively associated with PSA levels, whereas the other four loci also were associated with prostate cancer risk. We propose that a personalized PSA cutoff value, based on genotype, should be used when deciding to perform a prostate biopsy.

Ancestry-shift refinement mapping of the C6orf97-ESR1 breast cancer susceptibility locus.

We used an approach that we term ancestry-shift refinement mapping to investigate an association, originally discovered in a GWAS of a Chinese population, between rs2046210[T] and breast cancer susceptibility. The locus is on 6q25.1 in proximity to the C6orf97 and estrogen receptor alpha (ESR1) genes. We identified a panel of SNPs that are correlated with rs2046210 in Chinese, but not necessarily so in other ancestral populations, and genotyped them in breast cancer case:control samples of Asian, European, and African origin, a total of 10,176 cases and 13,286 controls. We found that rs2046210[T] does not confer substantial risk of breast cancer in Europeans and Africans (OR = 1.04, P = 0.099, and OR = 0.98, P = 0.77, respectively). Rather, in those ancestries, an association signal arises from a group of less common SNPs typified by rs9397435. The rs9397435[G] allele was found to confer risk of breast cancer in European (OR = 1.15, P = 1.2 x 10(-3)), African (OR = 1.35, P = 0.014), and Asian (OR = 1.23, P = 2.9 x 10(-4)) population samples. Combined over all ancestries, the OR was 1.19 (P = 3.9 x 10(-7)), was without significant heterogeneity between ancestries (P(het) = 0.36) and the SNP fully accounted for the association signal in each ancestry. Haplotypes bearing rs9397435[G] are well tagged by rs2046210[T] only in Asians. The rs9397435[G] allele showed associations with both estrogen receptor positive and estrogen receptor negative breast cancer. Using early-draft data from the 1,000 Genomes project, we found that the risk allele of a novel SNP (rs77275268), which is closely correlated with rs9397435, disrupts a partially methylated CpG sequence within a known CTCF binding site. These studies demonstrate that shifting the analysis among ancestral populations can provide valuable resolution in association mapping.

A sequence variant at 4p16.3 confers susceptibility to urinary bladder cancer.

Previously, we reported germline DNA variants associated with risk of urinary bladder cancer (UBC) in Dutch and Icelandic subjects. Here we expanded the Icelandic sample set and tested the top 20 markers from the combined analysis in several European case-control sample sets, with a total of 4,739 cases and 45,549 controls. The T allele of rs798766 on 4p16.3 was found to associate with UBC (odds ratio = 1.24, P = 9.9 x 10(-12)). rs798766 is located in an intron of TACC3, 70 kb from FGFR3, which often harbors activating somatic mutations in low-grade, noninvasive UBC. Notably, rs798766[T] shows stronger association with low-grade and low-stage UBC than with more aggressive forms of the disease and is associated with higher risk of recurrence in low-grade stage Ta tumors. The frequency of rs798766[T] is higher in Ta tumors that carry an activating mutation in FGFR3 than in Ta tumors with wild-type FGFR3. Our results show a link between germline variants, somatic mutations of FGFR3 and risk of UBC.

Parental origin of sequence variants associated with complex diseases.

Effects of susceptibility variants may depend on from which parent they are inherited. Although many associations between sequence variants and human traits have been discovered through genome-wide associations, the impact of parental origin has largely been ignored. Here we show that for 38,167 Icelanders genotyped using single nucleotide polymorphism (SNP) chips, the parental origin of most alleles can be determined. For this we used a combination of genealogy and long-range phasing. We then focused on SNPs that associate with diseases and are within 500 kilobases of known imprinted genes. Seven independent SNP associations were examined. Five-one with breast cancer, one with basal-cell carcinoma and three with type 2 diabetes-have parental-origin-specific associations. These variants are located in two genomic regions, 11p15 and 7q32, each harbouring a cluster of imprinted genes. Furthermore, we observed a novel association between the SNP rs2334499 at 11p15 and type 2 diabetes. Here the allele that confers risk when paternally inherited is protective when maternally transmitted. We identified a differentially methylated CTCF-binding site at 11p15 and demonstrated correlation of rs2334499 with decreased methylation of that site.

Whole genome association mapping by incompatibilities and local perfect phylogenies.

BACKGROUND: With current technology, vast amounts of data can be cheaply and efficiently produced in association studies, and to prevent data analysis to become the bottleneck of studies, fast and efficient analysis methods that scale to such data set sizes must be developed. RESULTS: We present a fast method for accurate localisation of disease causing variants in high density case-control association mapping experiments with large numbers of cases and controls. The method searches for significant clustering of case chromosomes in the "perfect" phylogenetic tree defined by the largest region around each marker that is compatible with a single phylogenetic tree. This perfect phylogenetic tree is treated as a decision tree for determining disease status, and scored by its accuracy as a decision tree. The rationale for this is that the perfect phylogeny near a disease affecting mutation should provide more information about the affected/unaffected classification than random trees. If regions of compatibility contain few markers, due to e.g. large marker spacing, the algorithm can allow the inclusion of incompatibility markers in order to enlarge the regions prior to estimating their phylogeny. Haplotype data and phased genotype data can be analysed. The power and efficiency of the method is investigated on 1) simulated genotype data under different models of disease determination 2) artificial data sets created from the HapMap ressource, and 3) data sets used for testing of other methods in order to compare with these. Our method has the same accuracy as single marker association (SMA) in the simplest case of a single disease causing mutation and a constant recombination rate. However, when it comes to more complex scenarios of mutation heterogeneity and more complex haplotype structure such as found in the HapMap data our method outperforms SMA as well as other fast, data mining approaches such as HapMiner and Haplotype Pattern Mining (HPM) despite being significantly faster. For unphased genotype data, an initial step of estimating the phase only slightly decreases the power of the method. The method was also found to accurately localise the known susceptibility variants in an empirical data set--the DeltaF508 mutation for cystic fibrosis--where the susceptibility variant is already known--and to find significant signals for association between the CYP2D6 gene and poor drug metabolism, although for this dataset the highest association score is about 60 kb from the CYP2D6 gene. CONCLUSION: Our method has been implemented in the Blossoc (BLOck aSSOCiation) software. Using Blossoc, genome wide chip-based surveys of 3 million SNPs in 1000 cases and 1000 controls can be analysed in less than two CPU hours.