Email updates

Keep up to date with the latest news and content from Genome Biology and BioMed Central.

Open Access Highly Accessed Research

The mutational landscape of chromatin regulatory factors across 4,623 tumor samples

Abel Gonzalez-Perez1, Alba Jene-Sanz1 and Nuria Lopez-Bigas12*

Author affiliations

1 Research Unit on Biomedical Informatics, Department of Experimental and Health Sciences, Universitat Pompeu Fabra, Dr. Aiguader 88, Barcelona, Spain

2 Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain

For all author emails, please log on.

Citation and License

Genome Biology 2013, 14:r106  doi:10.1186/gb-2013-14-9-r106


The electronic version of this article is the complete one and can be found online at: http://genomebiology.com/2013/14/9/r106


Received:6 July 2013
Accepted:24 September 2013
Published:24 September 2013

© 2013 Gonzalez-Perez et al.; licensee BioMed Central Ltd.

This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

Chromatin regulatory factors are emerging as important genes in cancer development and are regarded as interesting candidates for novel targets for cancer treatment. However, we lack a comprehensive understanding of the role of this group of genes in different cancer types.

Results

We have analyzed 4,623 tumor samples from thirteen anatomical sites to determine which chromatin regulatory factors are candidate drivers in these different sites. We identify 34 chromatin regulatory factors that are likely drivers in tumors from at least one site, all with relatively low mutational frequency. We also analyze the relative importance of mutations in this group of genes for the development of tumorigenesis in each site, and in different tumor types from the same site.

Conclusions

We find that, although tumors from all thirteen sites show mutations in likely driver chromatin regulatory factors, these are more prevalent in tumors arising from certain tissues. With the exception of hematopoietic, liver and kidney tumors, as a median, the mutated factors are less than one fifth of all mutated drivers across all sites analyzed. We also show that mutations in two of these genes, MLL and EP300, correlate with broad expression changes across cancer cell lines, thus presenting at least one mechanism through which these mutations could contribute to tumorigenesis in cells of the corresponding tissues.

Background

Highly conserved molecular mechanisms are responsible for maintaining genome integrity and tightly regulated gene expression, which is essential for cell survival. Those include the fine regulation of chromatin structure, mainly maintained through three distinct processes: the post-translational modification of histone tails, the replacement of core histones by histone variants, and the direct structural remodeling by ATP-dependent chromatin-remodeling enzymes [1]. The proteins that control this system, broadly referred to as chromatin regulatory factors (CRFs), contribute to the establishment of chromatin structures that modulate the expression of large gene sets, either by establishing more inaccessible regions or by placing histone marks that open the chromatin and allow the binding of other factors. These CRFs help to maintain cellular identity, and mutations in them (commonly called epimutations) often lead to a de-regulation of gene expression that may contribute to tumorigenesis [2]. CRFs are broadly classified in three main groups: histone tail modifiers (including histone acetyltransferases, histone deacetylases (HDACs), histone methyltransferases and histone demethylases, that deposit or remove acetyl or methyl groups, respectively); DNA methyltransferases (DNMTs) and putative demethylases (that affect cytosines at CpG islands); and ATP-dependent chromatin remodeling complexes (responsible for the repositioning of nucleosomes).

Until recently, DNMT proteins had not been found mutated in cancer [3], but DNMT3A, and later DNMT1 and DNMT3B, were reported as altered in patients with myelodysplastic syndromes and in acute monocytic leukemia, where their mutation status also predicted prognosis [4,5]. Mutations in ATP-dependent chromatin-remodeling complexes are recurrent in, amongst others, ovarian and clear cell renal cancers [2]. The regulation of the trimethylation of histone H3 at K27 mark (H3K27me3) by the Polycomb complex, a key component to maintain stem cell identity, is also frequently compromised in a variety of cancer types, including those from breast, bladder, pancreas, prostate and lymphomas [6]. Histone demethylases have also been implicated in the development of a wide variety of tumors. Moreover, recent whole exome sequencing studies in large cohorts of tumor samples have highlighted as main findings the inactivating mutations on proteins that regulate the epigenomic state of cells [7]. Alterations in KAT6B [8], SMARCC1 [9] and NSD1 [10] have been described in uterine, cervical and skin pre-malignant lesions, respectively. This presents these proteins as potential biomarkers, thus adding early cancer detection to the possible uses of CRFs in the clinic.

This current accumulation of evidence for the role of CRFs in cancer has attracted the attention of the scientific community towards CRFs as novel targets for cancer treatment. In 2006, the first HDAC inhibitor (HDACi), Vorinostat, was approved by the US Food and Drugs Administration to treat a specific type of lymphoma, and more than 20 molecules of this type are currently under preclinical and clinical investigation [11]. Some DNMT inhibitors have been recently approved by the US Food and Drugs Administration to treat myelodysplastic syndromes, and their combination with HDACi is a subject of intense study in clinical trials [12]. Some studies raise hopes for the possible use of HDACis to overcome drug resistance [13]. Interestingly, an in-depth review by Patel et al. on 46 potentially druggable yet chemically unexplored proteins in the Cancer Gene Census (CGC) identified six CRFs: ATRX, KAT6A, KDM6A, NSD3, PBRM1 and SMARCA4 [14].

Even though CRFs are emerging as important players in cancer development [15-20], to our knowledge no systematic analysis on the alterations of a comprehensive catalog of CRFs in different tumors has been performed to date. Moreover, most studies have focused their efforts in the in-depth characterization of specific genes that appear mutated at high frequencies, underestimating the impact of lowly recurrent drivers (those genes whose mutation is likely to be functional, but occurs in few samples) on tumorigenesis. For instance, a very recent report [21] focused only on the SWI/SNF family took into account the frequency of mutations of their members rather than their likelihood of driving tumorigenesis.

In this paper, we carry out a systematic exploration of the role of CRFs in tumorigenesis in different tissues. To that end, we first compiled and manually curated a comprehensive list of CRFs, for which we annotated any previously known implications in cancer. Secondly, we analyzed 4,623 tumor samples from 13 anatomical sites to identify which of the CRFs are driver candidates in these different sites, employing two approaches recently introduced by us [22,23]. Finally, we took advantage of the profiles of genomic and transcriptomic alterations revealed by the Cancer Cell Line Encyclopedia (CCLE) [24] to explore the effects of mutations in two likely driver CRFs on the expression of broad gene modules across 905 cancer cell lines.

Results

Analysis of chromatin regulatory factor tumor somatic mutations identifies 34 likely drivers in 13 cancer sites

To determine which CRFs may be involved in cancer emergence and development in primary tumors from 13 anatomical sites upon mutation, we first collected and manually curated a list of CRFs from the literature. This catalog contained 183 proteins grouped into eleven major functional classes, the most populated of which were the HDACs, the histone acetyltransferases and the histone methyltransferases. (The detailed list of CRFs in all functional classes is presented in Additional file 1: Table S1). Only 26 of them are included in the CGC. However, we found that many of these CRFs (115 out of 183) have some evidence, mainly in scattered reports from the past two years, of genomic or transcriptomic alterations in human tumors (Table  1 and Additional file 1: Table S2).

Additional file 1. Supplementary Tables S1, S2, S3 and S4 with titles and descriptions, and supplementary references

Format: PDF Size: 214KB Download file

This file can be viewed with: Adobe Acrobat ReaderOpen Data

Table 1. Described oncogenic alterations in chromatin regulatory factors that are candidate drivers in at least one tissue

In IntOGen-mutations [83], during the past year, we have collected and analyzed datasets of cancer somatic mutations produced by several research groups across the world. Some of them have been generated within the framework of large international initiatives like The Cancer Genome Atlas (TCGA) [84] and the International Cancer Genomes Consortium [85], while others are the fruit of independent laboratories. Taken together, these datasets [86] contain somatic mutations detected in 4,623 primary tumor samples obtained from 13 anatomical sites (Table  2). Each dataset has been analyzed separately, to compensate for differences between tumor histologies and subtypes, and between sequencing analysis pipelines. First, we used an approach recently developed by us, OncodriveFM [22], to detect genes that, across the cohort of tumor samples, tend to accumulate functional mutations. We give the name 'FM bias’ to this significant trend towards the accumulation of functional mutations. The FM bias is a signal of positive selection during cancer development and therefore FM-biased genes are likely candidates to drivers. Second, we identified genes whose mutations tend to significantly cluster in certain regions of their protein sequence (CLUST bias) also via an approach recently developed in our group, OncodriveCLUST [23]. Both FM-biased and CLUST-biased genes constitute sound candidates to cancer drivers [87] in these 13 anatomical sites. We have also combined the P values of FM bias and CLUST bias of individual genes across the datasets of tumor samples obtained from the same anatomical site. In summary, we have obtained a measurement of FM bias and CLUST bias for each mutated gene at the level of one dataset of tumor samples (or project), and also at the level of each anatomical site (or tissue). This catalog of likely driver genes has allowed us, for the first time, to systematically explore the involvement of epigenetic mechanisms (via mutations in CRFs) in tumorigenesis in 4,623 tumor samples from 13 anatomical sites.

Table 2. Description of the datasets of tumor somatic mutations collected and analyzed to detect candidate cancer driver genes

After an exhaustive search within the list of likely driver genes, we found that 34 CRFs from our manually curated list are FM biased and/or CLUST biased in at least one site (Figure  1, upper panel). Sixteen of them appear as likely drivers in more than one project, and only liver carcinomas appear free of likely driver CRFs - although this may be attributed to the small sample size of the dataset. Several driver CRFs are mutated at frequencies above 10% in at least one site (Figure  1, lower panel). Functional relationships among many of them - 124 CRF genes can be mapped onto a pre-compiled [108] functional interaction network (Figure  2) - suggest the possibility that mutations in different genes produce similar malignancies (see below). We can therefore make the general observation that CRFs - 34 in the dataset collected by us - potentially act as mutational drivers in most of the cancer sites studied.

thumbnailFigure 1. Likely driver chromatin regulatory factors across the datasets of somatic mutations in IntOGen-mutations. The heat-map in the top panel identifies FM-biased and CLUST-biased CRFs across the 31 datasets from 13 sites in IntOGen-mutations, whose original projects are detailed in the middle panel. The heat-map in the bottom panel contains the number of samples with mutations in each likely driver CRF in each site. Cells in the heat-map are colored following mutational frequency.

thumbnailFigure 2. Chromatin regulatory factors within their context of functional interactions. Network of functional interactions among CRFs mapped to the Cytoscape FI plugin network. Square nodes represent likely driver CRFs, circle nodes other CRFs within the catalog, and diamond nodes represent linker genes. CRFs functions are color-coded, and genes in the same complex are grouped and circled.

Driver candidates are significantly overrepresented within our catalog of CRFs (34 driver CRFs from 183 human CRFs in our list versus 348 total drivers from 22,696 human genes; Fisher’s P value 1.26 × 10-25). In addition, when analyzed as a group, the 183 CRFs in our catalog appear FM biased in all sites except liver (Figure  3A), which indicates that collectively they tend to accumulate mutations that on average possess higher functional impact than the background of the corresponding tumors. Taken together, these two observations suggest that CRFs as a group may have an important role in tumorigenesis in the 13 sites with data in IntOGen.

thumbnailFigure 3. FM bias, mutation frequencies and mutually exclusivity of chromatin regulatory factors as part of complexes. (A) Left heat-map shows the P value of FM bias analysis for all CRFs and for each complex. Right heat-map shows the number of samples with PAMs in the complex and the color indicates the mutation frequency (number of samples with PAMs divided by number of samples of this cancer type analyzed). (B) Heat-map of samples and genes of each complex, PAMs are represented as green cells in the heat-map. Tumor samples from each site have headers of the corresponding colors. Samples and genes in the heat-map are ordered based on mutually exclusive alterations within each site using Gitools built-in function for this purpose. Number of samples with PAMs in the gene (N) and the mutation frequency (Freq) of the gene in whole dataset are shown at the right of each heat-map. Gene names in bold indicated that the gene is one of the 34 detected as candidate drivers. PAM, protein-affecting mutations.

Because CRFs usually act as multiprotein complexes, we also determined whether the best established of these complexes exhibit discernible signals of positive selection as a group across tumor samples. We did this in two ways. First, we computed the FM bias of six complexes described in Additional file 1: Table S1 and whose components appear illustrated in the network of functional interactions in Figure  2. We established that five of the complexes - ISWI being the exception - significantly accumulate highly impacting mutations in at least one site (Figure  3A). Second, we observed that pairs of proteins of the same complex tend to be mutated following a pattern of mutual exclusivity within cancer sites (Figure  3B and Additional file 1: Table S3). For example, the exploration of the SWI/SNF complex in breast tumors revealed that ARID1A tends to be mutated in samples where neither SMARCA4, ARID2 nor SMARCA2 are mutated. These two observations imply that multiprotein complexes, rather that individual genes, are the subjects of positive selection during tumorigenesis in the cancer sites under study.

The implication of chromatin regulatory factors in tumorigenesis strongly depends on the anatomical site and the tumor type

To determine whether there are differences in the implication of CRFs in tumorigenesis across the anatomical sites in IntOGen, we first computed the number of likely driver genes in general, and likely driver CRFs in particular, that bear protein sequence-affecting mutations, or PAMs (non-synonymous, stop, frameshift-causing insertions or deletions (indels)) in each tumor sample. From these data, the simplest way of representing the relative importance of mutations in CRFs in tumorigenesis across sites consists of counting the number of samples with at least one FM-biased CRF bearing a PAM (Figure  4A). In this metric, bladder urothelial carcinomas and endometrial carcinomas stand out, with more than 80% and 60%, respectively, of the samples with at least one mutated CRF. On the opposite extreme, less than 10% of brain and hematopoietic tumor samples contain mutated likely driver CRFs.

thumbnailFigure 4. Relative importance of chromatin regulatory factors in tumorigenesis across sites. (A) Histograms of the fraction of samples with 0 (green) or at least one (red) likely driver CRF with PAMs in each site. (B) Boxplots representing the distribution of fraction of CRFs with PAMs with respect to all FM-biased genes with PAMs in each sample (CF ratios) of samples from each site with at least one mutation in a CRF (red fraction in panel A). (C) Boxplots representing the distribution of CF ratios of samples from each of the three projects focused on brain tumors. CRF, chromatin regulatory factors; DKFZ, German Cancer Research Center; JHU, Johns Hopkins University; TCGA, The Cancer Genome Atlas.

We then computed the fraction of CRFs with PAMs with respect to all FM-biased genes with PAMs in each sample (CF ratio) (Figure  4B). This measure gives an indication of the relative importance of CRFs in the tumorigenesis process in each sample. Although liver or hematopoietic are not among the sites with the highest proportion of tumor samples with mutated CRFs (Figure  4A), these appear to be very important in the development of tumors in these sites (see the corresponding boxplots of Figure  4B). A closer look at the repertoire of mutated drivers in the samples of the three brain tumor datasets currently in IntOGen reveals that whereas mutations in classic tumor suppressors and oncogenes dominate the landscape of glioblastomas, mutations in CRFs are more predominant in pediatric medulloblastomas. The median of the ratio of mutations in CRFs to mutations in all drivers across medulloblastoma samples is 0.4, compared to 0.21 and 0.1 in glioblastoma JHU (Johns Hopkins University; see Table  2) and glioblastoma TCGA (The Cancer Genome Atlas; see Table  2), respectively (Figure  4C). The samples of these two glioblastoma datasets exhibit a repertoire of mutated 'classical’ tumor suppressors and oncogenes, such as TP53, PTEN and EGFR (Figure  5). As observed in the previous section, mutations in CRFs are likely drivers in tumors from most cancer types. Nevertheless, the latter result suggests that these mutations are circumscribed to a relatively small number of tumor samples, although future reviews of the catalogs of CRFs may increase the proportions calculated here.

thumbnailFigure 5. Mutational status of tumor samples from the three brain datasets included in IntOGen. The genes represented in the heat-map comprise all FM-biased CRFs that bear one mutation in at least one brain tumor sample (in bold typeface) plus the top 15 FM-biased genes in brain obtained from IntOGen. Mutations are represented by their MutationAssesor [109] functional impact scores (FIS). Samples and genes in the heat-map are ordered based on mutually exclusive alterations within dataset. FIS, functional impact score; MA, MutationAssessor score. JHU, Johns Hopkins University; TCGA, The Cancer Genome Atlas.

Mutations in chromatin regulatory factors correlate with transcriptomic alterations of gene modules in cancer cell lines

To further understand the possible implication of CRFs in tumorigenesis, we explored the effects of CRF mutations on changes in the transcriptional levels of broad gene sets in cancer cell lines. To this end, we employed the data produced by The Cancer Cell Line Encyclopedia project, which has sequenced 1,651 protein-coding genes, of which 43 are CRFs according to our curated list (see Additional file 1: Table S1 for a detailed classification). First, to check whether cancer cell lines behave comparably to primary tumors in the patterns of altered gene modules, we carried out a sample level enrichment analysis (SLEA) [110] over cancer cell lines using Gene Ontology Biological Process terms that are altered in specific cancer tissues. We found that cancer cell lines repeated the transcriptional profiles typical of their corresponding primary tumors (Additional file 2: Figure S1).

We then assessed the transcriptional impact of PAMs on EP300 and MLL3 (the only CRFs sustaining PAMs in sufficient cell lines: 115 and 191, respectively) to determine whether the impact of these PAMs on epigenetic regulation could translate into changes of the transcriptional levels of broad gene sets. The underlying hypothesis was that genes whose transcription was modulated by specific histone marks that became affected by PAMs on these two genes would present expression changes detectable when analyzed as a group. We collected regulatory modules of histone modifications in three cell types (Additional file 1: Table S4) and performed SLEA separately on cell lines originated from blood and solid tissues (Figure  6). As a result of the SLEA, we obtained a value of significance of the over-expression or under-expression (as a z-score) of each module in each cell line. We then compared the z-scores of cell lines that bear mutations in the gene in question (EP300 or MLL3) to those cell lines where it does not, using the Wilcoxon-Mann–Whitney test. The P-values of the right-tail and left-tail comparisons were then adjusted using the Benjamini-Hochberg approach. Figure  6 presents the modules that rendered either significant right-tail or left-tail P values for any of the two genes. It shows that, in general, cell lines from solid tissues with mutations in either EP300 or MLL3 exhibited lower expression of repressed chromatin gene modules (H3K27me3 and late replicating genes), and higher expression of gene modules with activating histone marks (marked by H3K4me3 and H3K9ac; Table  1). The under-expression of the H3K27me3 module, regulated by Polycomb, has been associated to a stem cell-like signature and more aggressive tumors [86]. Moreover, cell lines with mutations in MLL3 showed higher expression of cell cycle-related modules. Taken together, these results suggest that mutations in CRFs may affect the transcriptional levels of gene sets bearing histone marks related to these CRFs.

Additional file 2. Supplementary Figure S1 and S2.

Format: PDF Size: 831KB Download file

This file can be viewed with: Adobe Acrobat ReaderOpen Data

thumbnailFigure 6. Effect of PAMs in EP300 and MLL3 on the transcription of broad gene modules across cancer cell lines. Cancer cell lines originated from solid tissues (Additional file 2: Figure S1) are enriched (SLEA) for regulatory modules (Additional file 1: Table S4) and selected pathways from Kyoto Encyclopedia of Genes and Genomes. The first two panels in both A and B correspond to mean enrichment z-scores in wild type and mutant cell lines. The difference between the two enrichment groups, assessed through a Wilcoxon-Mann–Whitney group comparison test, is indicated at the right. (A)EP300 mutation status. (B)MLL3 mutation status.

Discussion

In this study, we found that several CRFs are likely involved in tumorigenesis in cancers from 13 anatomical sites. We uncovered these genes as putative causes of the studied malignancies through the use of the FM bias and CLUST bias analyses, rather than the mere recurrence of mutations in genes across tumor samples. Moreover, by focusing on multiprotein complexes formed by several CRFs, we found evidence that suggest that these, rather than individual genes, are the subjects of positive selection during tumorigenesis. These two approaches constitute novelties with respect to the most recent and comprehensive analysis [21], which found recurrent mutations in SWI/SNF proteins across more than 650 tumor samples of 10 anatomical sites. Another important methodological novelty of our work consists in the use of CF ratios to assess the relevance of mutations in CRFs in tumorigenesis in cancers from different sites. The employment of this ratio normalizes the number of CRFs with PAMs in the samples of a site by its intrinsic burden of mutations in driver genes. It is thus possible to observe that PAMs in CRFs, although highly prevalent in carcinomas of the uterus, probably play a relatively small role in their tumorigenesis because these tumors bear mutations in many other driver genes. However, mutations in CRFs appear to play a bigger role in tumorigenesis in hematopoietic malignancies than they do in tumors from other sites, although only few hematopoietic tumors bear PAMs in CRFs (Figure  4).

A group of pediatric medulloblastomas also possess abnormally high CF ratios, which implies that a high proportion of their mutated drivers are actually CRFs. It has been suggested that both pediatric and hematopoietic malignancies have very low mutational rates and therefore fewer drivers take part in their emergence than in solid adult tumors [111]. One could hypothesize from our results that alteration of either the transcriptional control or the chromatin maintenance of broad gene modules - as we observed in cell lines - via mutations in CRFs may be the crucial step of tumorigenesis in at least some of these tumors. This hypothesis, which could be experimentally tested, is another important contribution of the present work.

A third important contribution is the list of putative driver CRFs, which is available at IntOGen [112]. In particular, two of them were uncovered as putative drivers in more than one site (CHD4 and ATF7IP) and are not annotated in the CGC [82]. They therefore constitute interesting candidates for novel epigenetic drivers (Figure  1). These additions to the list of driver CRFs might contribute to the research for anticancer drugs that takes CRFs as suitable targets.

Conclusions

We present the first systematic approach to characterize the repertoire of CRFs that could constitute mutational cancer drivers in tumors from 13 anatomical sites. We found that likely driver CRFs appear across tumor samples from most of these 13 sites, although the number of affected samples is in general low, except in the case of tumors from several sites, such as bladder, kidney and uterus. Mutations in CRFs appear to be in general only one of several contributing mechanisms towards tumorigenesis in most cancer samples. Finally, we have proved that mutations in two CRFs correlate with broad expression changes across cancer cell lines, thus presenting at least one mechanism through which these mutations could contribute to tumorigenesis in cells of the corresponding tissues. Our results expand the current knowledge on the involvement of CRFs in tumorigenesis in several malignancies. Furthermore, they can contribute to formulate hypotheses on the mechanistic basis for this association. All the results presented here are available for browsing through the IntOGen-mutations platform [83,112] and using Gitools interactive heat-maps [113].

Materials and methods

Chromatin regulatory factors

We manually compiled a list of 183 genes coding for CRF proteins from the literature, based on protein function and known essential association to complexes important for the regulation of chromatin structure. A detailed classification of these CRFs is presented in Additional file 1: Table S1; the information was obtained from the Uniprot database [114] and the manuscripts referenced within the additional file. The relevant proteins for the purpose of this analysis are described in Table  1 and Additional file 1: Table S2.

FM-biased genes in primary tumors

FM-biased genes exhibit a bias towards the accumulation of functional mutations across a cohort of tumor samples and are therefore candidate cancer drivers. We have compiled 31 datasets of tumors from 13 anatomical sites and detected the FM-biased genes in each of them with the approach described in [22]. Genes that were not expressed across the major (TCGA) datasets included in IntOGen (obtained from syn1734155) were eliminated from the OncodriveFM analysis at this point. The overlap of drivers obtained from different datasets of mutations detected in tumors from the same anatomical site is shown in Additional file 2: Figure S2. Finally, we combined the gene-wise P values obtained for datasets of the same anatomical site to obtain a single P value that measures the bias of the gene towards the accumulation of functional mutations in different tumors from the same site. The corrected genes FM bias P values in these 13 tissues are stored in the IntOGen knowledgebase [87]. The collection of the datasets of tumor somatic mutations, their processing and browsing through IntOGen are thoroughly described in [83]. Details of the 31 tumor somatic mutations datasets are presented in Table  2.

CLUST-biased genes in primary tumors

PAMs in CLUST-biased genes tend to be grouped in regions of the proteins in a higher degree than synonymous mutations across the same dataset. This grouping constitutes another signal of positive selection that points to likely cancer drivers. The method to compute the CLUST bias in genes across datasets of tumor somatic mutations (OncodriveCLUST) is described in Tamborero et al., [23]. We computed the CLUST bias of all genes with PAMs across the 31 datasets compiled and stored in IntOGen-mutations [83]. Genes that were not expressed across the major (TCGA) datasets included in IntOGen (obtained from syn1734155) were eliminated from the OncodriveCLUST analysis at this point. As with the FM bias, we combined the gene-wise P values obtained for datasets of tumor samples from the same anatomical site. The corrected genes’ CLUST bias P values in these 13 tissues are stored in the IntOGen knowledgebase [87]. The collection of the datasets of tumor somatic mutations, their processing and browsing through IntOGen are thoroughly described in [83]. Details of the 31 tumor somatic mutations datasets are presented in Table  2.

Analysis of mutational frequencies of tumor samples

We defined a group of broad consequence types as corresponding to PAMs for all analyses of the mutational frequencies of tumor samples. All non-synonymous, stop and frameshift indels were included in this group. We recorded two numbers in the 4,623 tumor samples included in the study: the number of PAMs in any of the 34 likely driver CRFs detected across the 13 sites; and the number of PAMs in any of the 382 likely driver genes detected across the 13 sites. We then computed the ratio (CRFs-to-drivers ratio, or CF ratio) between these two numbers to assess the relevance of mutations in CRFs in tumorigenesis in every tumor sample. Note that because the 34 likely driver CRFs were included within the catalog of 382 likely drivers, the CF ratio takes values between 0 (no mutations in CRFs) and 1 (all mutated drivers in the sample are CRFs). Finally, we computed the number of tumor samples from each site with at least one PAM in a CRF and the distribution of their CF ratios.

Functional network analysis

We mapped the 183 CRFs in our catalog to the functional interactions network within the Cytoscape FI plugin [108,115], allowing the presence of linker genes to maximize the number of connected CRFs. Using Cytoscape, we then grouped genes in the same multiprotein complex (from the ones shown in Additional file 1: Table S2). We also mapped the biological functions of CRFs in the network using nodes colors, and whether they appeared as likely drivers through nodes shapes.

Cancer cell lines data processing

Expression arrays from the CCLE were downloaded from the Gene Expression Omnibus [GEO:GSE36133] as raw CEL files, and pre-processed as previously described [110]. The input data for enrichment analysis was obtained by median centering the expression value of each gene across cancer cell lines and dividing this value by the standard deviation. The obtained value is the measure of expression level for the gene in a sample as compared to its expression level in all other samples in the dataset. We built separate expression matrices for cancer cell lines obtained from hematological system or solid primary cells, since the expression profiles of these two groups were shown to clearly differ in the original publication [24].

SLEA was performed using Gitools version 1.6.0 [116]. We used the z-score method as described previously [117]. This method compares the mean (or median) expression value of genes in each module to a distribution of mean (or median) of 10,000 random modules of the same size. Such enrichment analysis is run for each sample and the result is a z-score, which is a measure of the difference between the observed and expected mean (or median) expression values for genes in a module. We applied the mean z-score enrichment values, which are the arithmetic means of z-scores for individual samples, separately in cell lines obtained from hematological system or in those obtained from solid primary cells. To test for significant differences between the z-score means between groups of cell lines we used the Mann–Whitney test [118] implemented in Gitools. All heat-maps were generated with Gitools [119].

To detect potential PAMs in genes within the list of CRFs (Additional file 1: Table S1), we downloaded processed mutations data (single nucleotide variants and small indels) for 1,651 protein-coding genes (7 May 2012 version, excluding common polymorphisms and single nucleotide variants with an allelic fraction >10%) from the CCLE website [120]. We computed the consequence types of these variants using the Ensembl (v69) Variant Effect Predictor wrapped within the IntOGen-mutations pipeline [83].

Public gene regulation datasets

We collected lists of genes occupied by a specific histone mark or bound by a regulatory factor, and computationally predicted chromatin states, from available sources (Additional file 1: Table S4). These included human genome-wide occupancy datasets from ChIP-seq experiments in several cell types [121-125] that we processed using Bowtie [126] (version 0.12.5, hg19 genome assembly, unique alignments, allowing two mismatches) for short read aligning. For peak detection of transcription factors we used MACS [127] (version 1.4.1, settings: --nomodel and --bw parameter set to twice the shift size whenever a control immunoprecipitation was not available). For broad histone modifications (that is, H3K27me3), we used SICER [128] (version 1.1, setting gap size to 600). Regions were assigned to protein-coding genes (Ensembl v69) if they overlapped either to the gene body or up to 5 kb upstream from the transcription start site, using BedTools [129]. Overall peak calling performance was evaluated with CEAS [130]. Other gene sets were obtained from KEGG [131] and Gene Ontology [132]. The list and mappings (in Ensembl v67 IDs) of KEGG and Gene Ontology Biological Process terms were downloaded through the Gitools importer [116].

Abbreviations

CCLE: Cancer cell line encyclopedia; CF: Mutations in CRFs-to-mutations in drivers ratio; CGC: Cancer gene census; CRF: Chromatin regulatory factors; DCC: ICGC Data coordination center; DKFZ: German cancer research center; DNMT: DNA methyltransferases; H3K27me3: Trimethylation of histone H3 at K27 mark; HDAC: Histone deacetylases; HDACi: Histone deacetylase inhibitor; IACR: International Agency for Research on Cancer; ICGC: International cancer genomes consortium; JHU: Johns Hopkins University; PAM: Protein-affecting mutation; SLEA: Sample level enrichment analysis; SM: Supplementary Material of articles; TCGA: The cancer genome Atlas.

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

The three authors designed the study. AJ-S curated the list of CRFs and performed SLEA and analyses of mutations in cell lines. AG-P analyzed mutation data in 4,623 primary tumors and identified FM-biased and CLUST-biased genes in different tissues. NL-B supervised the study. The three authors contributed to drafting and editing the manuscript. All authors read and approved the final manuscript.

Acknowledgements

We acknowledge funding from the Spanish Ministry of Economy and Competitivity (grant numbers SAF2009-06954 and SAF2012-36199) and the Spanish National Institute of Bioinformatics. AJ-S is supported by an FPI fellowship. We gratefully acknowledge contributions from the TCGA Research Network and its TCGA Pan-Cancer Analysis Working Group (contributing consortium members are listed in Additional file 3: Table S5). The TCGA Pan-Cancer Analysis Working Group is coordinated by JM Stuart, C Sander and I Shmulevich. We are also grateful to the ICGC for the tumor genome resequencing data generated.

Additional file 3. Supplementary Table S5.

Format: XLSX Size: 34KB Download fileOpen Data

References

  1. Papamichos-Chronakis M, Peterson CL: Chromatin and the genome integrity network.

    Nat Rev Genet 2013, 14:62-75. OpenURL

  2. Elsässer SJ, Allis CD, Lewis PW: New epigenetic drivers of cancers.

    Science 2011, 331:1145-1146. OpenURL

  3. Bestor TH: Unanswered questions about the role of promoter methylation in carcinogenesis.

    Ann N Y Acad Sci 2003, 983:22-27. OpenURL

  4. Yan X-J, Xu J, Gu Z-H, Pan C-M, Lu G, Shen Y, Shi JY, Zhu YM, Tang L, Zhang XW, Liang WX, Mi JQ, Song HD, Li KQ, Chen Z, Chen SJ: Exome sequencing identifies somatic mutations of DNA methyltransferase gene DNMT3A in acute monocytic leukemia.

    Nat Genet 2011, 43:309-315. OpenURL

  5. Walter MJ, Ding L, Shen D, Shao J, Grillot M, McLellan M, Fulton R, Schmidt H, Kalicki-Veizer J, O'Laughlin M, Kandoth C, Baty J, Westervelt P, DiPersio JF, Mardis ER, Wilson RK, Ley TJ, Graubert TA: Recurrent DNMT3A mutations in patients with myelodysplastic syndromes.

    Leukemia 2011, 25:1153-1158. OpenURL

  6. Jene-Sanz A, Váraljai R, Vilkova AV, Khramtsova GF, Khramtsov AI, Olopade OI, Lopez-Bigas N, Benevolenskaya EV: Expression of polycomb targets predicts breast cancer prognosis.

    Mol Cell Biol 2013, 33:3951-3961. OpenURL

  7. You JS, Jones PA: Cancer genetics and epigenetics: two sides of the same coin?

    Cancer Cell 2012, 22:9-20. OpenURL

  8. Moore SDP, Herrick SR, Ince TA, Kleinman MS, Cin PD, Morton CC, Quade BJ: Uterine leiomyomata with t(10;17) disrupt the histone acetyltransferase MORF.

    Cancer Res 2004, 64:5570-5577. OpenURL

  9. Shadeo A, Chari R, Lonergan KM, Pusic A, Miller D, Ehlen T, Van Niekerk D, Matisic J, Richards-Kortum R, Follen M, Guillaud M, Lam WL, MacAulay C: Up regulation in gene expression of chromatin remodelling factors in cervical intraepithelial neoplasia.

    BMC Genomics 2008, 9:64. OpenURL

  10. Quintana RM, Dupuy AJ, Bravo A, Casanova ML, Alameda JP, Page A, Sánchez-Viera M, Ramírez A, Navarro M: A transposon-based analysis of gene mutations related to skin cancer development.

    J Invest Dermatol 2013, 133:239-248. OpenURL

  11. Giannini G, Cabri W, Fattorusso C, Rodriquez M: Histone deacetylase inhibitors in the treatment of cancer: overview and perspectives.

    Future Med Chem 2012, 4:1439-1460. OpenURL

  12. Baylin SB, Jones PA: A decade of exploring the cancer epigenome — biological and translational implications.

    Nat Rev Cancer 2011, 11:726-734. OpenURL

  13. Sharma SV, Haber DA, Settleman J: Cell line-based platforms to evaluate the therapeutic efficacy of candidate anticancer agents.

    Nat Rev Cancer 2010, 10:241-253. OpenURL

  14. Patel MN, Halling-Brown MD, Tym JE, Workman P, Al-Lazikani B: Objective assessment of cancer genes for drug discovery.

    Nat Rev Drug Discov 2013, 12:35-50. OpenURL

  15. Ley TJ, Ding L, Walter MJ, McLellan MD, Lamprecht T, Larson DE, Kandoth C, Payton JE, Baty J, Welch J, Harris CC, Lichti CF, Townsend RR, Fulton RS, Dooling DJ, Koboldt DC, Schmidt H, Zhang Q, Osborne JR, Lin L, O'Laughlin M, McMichael JF, Delehaunty KD, McGrath SD, Fulton LA, Magrini VJ, Vickery TL, Hundal J, Cook LL, Conyers JJ, et al.: DNMT3A mutations in acute myeloid leukemia.

    New Engl J Med 2010, 363:2424-2433. OpenURL

  16. Yamashita Y, Yuan J, Suetake I, Suzuki H, Ishikawa Y, Choi YL, Ueno T, Soda M, Hamada T, Haruta H, Takada S, Miyazaki Y, Kiyoi H, Ito E, Naoe T, Tomonaga M, Toyota M, Tajima S, Iwama A, Mano H: Array-based genomic resequencing of human leukemia.

    Oncogene 2010, 29:3723-3731. OpenURL

  17. Uno K, Takita J, Yokomori K, Tanaka Y, Ohta S, Shimada H, Gilles FH, Sugita K, Abe S, Sako M, Hashizume K, Hayashi Y: Aberrations of the hSNF5/INI1 gene are restricted to malignant rhabdoid tumors or atypical teratoid/rhabdoid tumors in pediatric solid tumors.

    Genes Chromosomes Cancer 2002, 34:33-41. OpenURL

  18. Jiao Y, Shi C, Edil BH, de Wilde RF, Klimstra DS, Maitra A, Schulick RD, Tang LH, Wolfgang CL, Choti MA, Velculescu VE, Diaz LA Jr, Vogelstein B, Kinzler KW, Hruban RH, Papadopoulos N: DAXX/ATRX, MEN1, and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors.

    Science 2011, 331:1199-1203. OpenURL

  19. Banine F, Bartlett C, Gunawardena R, Muchardt C, Yaniv M, Knudsen ES, Weissman BE, Sherman LS: SWI/SNF chromatin-remodeling factors induce changes in DNA methylation to promote transcriptional activation.

    Cancer Res 2005, 65:3542-3547. OpenURL

  20. Jones S, Wang T-L, Shih I-M, Mao T-L, Nakayama K, Roden R, Glas R, Slamon D, Diaz LA Jr, Vogelstein B, Kinzler KW, Velculescu VE, Papadopoulos N: Frequent mutations of chromatin remodeling gene ARID1A in ovarian clear cell carcinoma.

    Science 2010, 330:228-231. OpenURL

  21. Shain AH, Pollack JR: The Spectrum of SWI/SNF mutations. Ubiquitous in human cancers.

    PLoS One 2013, 8:e55119. OpenURL

  22. Gonzalez-Perez A, Lopez-Bigas N: Functional impact bias reveals cancer drivers.

    Nucl Acids Res 2012, 40:e169. OpenURL

  23. Tamborero D, Gonzalez-Perez A, Lopez-Bigas N: Oncodrive CLUST: exploiting the positional clustering of somatic mutations to identify cancer genes.

    Bioinformatics 2013, 29:2238-2244. OpenURL

  24. Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, Wilson CJ, Lehár J, Kryukov GV, Sonkin D, Reddy A, Liu M, Murray L, Berger MF, Monahan JE, Morais P, Meltzer J, Korejwa A, Jané-Valbuena J, Mapa FA, Thibault J, Bric-Furlong E, Raman P, Shipway A, Engels IH, Cheng J, Yu GK, Yu J, Aspesi P Jr, de Silva M, Jagtap K, et al.: The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity.

    Nature 2012, 483:307-603. OpenURL

  25. Gui Y, Guo G, Huang Y, Hu X, Tang A, Gao S, Wu R, Chen C, Li X, Zhou L, He M, Li Z, Sun X, Jia W, Chen J, Yang S, Zhou F, Zhao X, Wan S, Ye R, Liang C, Liu Z, Huang P, Liu C, Jiang H, Wang Y, Zheng H, Sun L, Liu X, Jiang Z, et al.: Frequent mutations of chromatin remodeling genes in transitional cell carcinoma of the bladder.

    Nat Genet 2011, 43:875-878. OpenURL

  26. Fujimoto A, Totoki Y, Abe T, Boroevich KA, Hosoda F, Nguyen HH, Aoki M, Hosono N, Kubo M, Miya F, Arai Y, Takahashi H, Shirakihara T, Nagasaki M, Shibuya T, Nakano K, Watanabe-Makino K, Tanaka H, Nakamura H, Kusuda J, Ojima H, Shimada K, Okusaka T, Ueno M, Shigekawa Y, Kawakami Y, Arihiro K, Ohdan H, Gotoh K, Ishikawa O, et al.: Whole-genome sequencing of liver cancers identifies etiological influences on mutation patterns and recurrent mutations in chromatin regulators.

    Nat Genet 2012, 44:760-764. OpenURL

  27. Gallo ML, O’Hara AJ, Rudd ML, Urick ME, Hansen NF, O’Neil NJ, Price JC, Zhang S, England BM, Godwin AK, Sgroi DC, Hieter P, Mullikin JC, Merino MJ, Bell DW, NIH Intramural Sequencing Center (NISC) Comparative Sequencing Program: Exome sequencing of serous endometrial tumors identifies recurrent somatic mutations in chromatin-remodeling and ubiquitin ligase complex genes.

    Nat Genet 2012, 44:1310-1315. OpenURL

  28. The Cancer Genome Atlas Research Network: Comprehensive molecular characterization of human colon and rectal cancer.

    Nature 2012, 487:330-337. OpenURL

  29. Zang ZJ, Cutcutache I, Poon SL, Zhang SL, McPherson JR, Tao J, Rajasegaran V, Heng HL, Deng N, Gan A, Lim KH, Ong CK, Huang D, Chin SY, Tan IB, Ng CC, Yu W, Wu Y, Lee M, Wu J, Poh D, Wan WK, Rha SY, So J, Salto-Tellez M, Yeoh KG, Wong WK, Zhu YJ, Futreal PA, Pang B, et al.: Exome sequencing of gastric adenocarcinoma identifies recurrent somatic mutations in cell adhesion and chromatin remodeling genes.

    Nat Genet 2012, 44:570-574. OpenURL

  30. Biankin AV, Waddell N, Kassahn KS, Gingras M-C, Muthuswamy LB, Johns AL, Miller DK, Wilson PJ, Patch AM, Wu J, Chang DK, Cowley MJ, Gardiner BB, Song S, Harliwong I, Idrisoglu S, Nourse C, Nourbakhsh E, Manning S, Wani S, Gongora M, Pajic M, Scarlett CJ, Gill AJ, Pinho AV, Rooman I, Anderson M, Holmes O, Leonard C, Taylor D, et al.: Pancreatic cancer genomes reveal aberrations in axon guidance pathway genes.

    Nature 2012, 491:399-405. OpenURL

  31. Seo J-S, Ju YS, Lee W-C, Shin J-Y, Lee JK, Bleazard T, Lee J, Jung YJ, Kim JO, Shin JY, Yu SB, Kim J, Lee ER, Kang CH, Park IK, Rhee H, Lee SH, Kim JI, Kang JH, Kim YT: The transcriptional landscape and mutational profile of lung adenocarcinoma.

    Genome Res 2012, 22:2109-2119. OpenURL

  32. Love C, Sun Z, Jima D, Li G, Zhang J, Miles R, et al.: The genetic landscape of mutations in Burkitt lymphoma.

    Nat Genet 2012, 44:1321-1325. OpenURL

  33. Sausen M, Leary RJ, Jones S, Wu J, Reynolds CP, Liu X, et al.: Integrated genomic analyses identify ARID1A and ARID1B alterations in the childhood cancer neuroblastoma.

    Nat Genet 2013, 45:12-17. OpenURL

  34. Mamo A, Cavallone L, Tuzmen S, Chabot C, Ferrario C, Hassan S, et al.: An integrated genomic approach identifies ARID1A as a candidate tumor-suppressor gene in breast cancer.

    Oncogene 2012, 31:2090-2100. OpenURL

  35. Lindberg J, Mills IG, Klevebring D, Liu W, Neiman M, Xu J, et al.: The mitochondrial and autosomal mutation landscapes of prostate cancer.

    Eur Urol 2012, 63:702-708. OpenURL

  36. Watanabe Y, Castoro RJ, Kim HS, North B, Oikawa R, Hiraishi T, et al.: Frequent alteration of MLL3 Frameshift mutations in microsatellite deficient colorectal cancer.

    PLoS One 2011, 6:e23320. OpenURL

  37. Liu P, Morrison C, Wang L, Xiong D, Vedell P, Cui P, et al.: Identification of somatic mutations in non-small cell lung carcinomas using whole-exome sequencing.

    Carcinogenesis 2012, 33:1270-1276. OpenURL

  38. Ellis MJ, Ding L, Shen D, Luo J, Suman VJ, Wallis JW, et al.: Whole-genome analysis informs breast cancer response to aromatase inhibition.

    Nature 2012, 486:353-360. OpenURL

  39. Ruault M, Brun ME, Ventura M, Roizès G, De Sario A: MLL3, a new human member of the TRX/MLL gene family, maps to 7q36, a chromosome region frequently deleted in myeloid leukaemia.

    Gene 2002, 284:73-81. OpenURL

  40. Kim MS, Kim YR, Yoo NJ, Lee SH: Mutational analysis of DNMT3A gene in acute leukemias and common solid cancers.

    APMIS 2012, 12:85-94. OpenURL

  41. Bai X, Song Z, Fu Y, Yu Z, Zhao L, Zhao H, et al.: Clinicopathological significance and prognostic value of DNA methyltransferase 1, 3a, and 3b expressions in sporadic epithelial ovarian cancer.

    PLoS One 2012, 7:e40024. OpenURL

  42. Liu J, Lee W, Jiang Z, Chen Z, Jhunjhunwala S, Haverty PM, et al.: Genome and transcriptome sequencing of lung cancers reveal diverse mutational and splicing events.

    Genome Res 2012, 22:2315-2327. OpenURL

  43. Robinson G, Parker M, Kranenburg TA, Lu C, Chen X, Ding L, et al.: Novel mutations target distinct subgroups of medulloblastoma.

    Nature 2012, 488:43-48. OpenURL

  44. van Haaften G, Dalgliesh GL, Davies H, Chen L, Bignell G, Greenman C, et al.: Somatic mutations of the histone H3K27 demethylase gene UTX in human cancer.

    Nat Genet 2009, 41:521-523. OpenURL

  45. Paolicchi E, Crea F, Farrar WL, Green JE, Danesi R: Histone lysine demethylases in breast cancer.

    Cr Rev Oncol-Hem 2012, 86:97-103. OpenURL

  46. Liu L, Zhao E, Li C, Huang L, Xiao L, Cheng L, et al.: TRIM28, a new molecular marker predicting metastasis and survival in early-stage non-small cell lung cancer.

    Cancer Epidem 2013, 37:71-78. OpenURL

  47. Shain AH, Giacomini CP, Matsukuma K, Karikari CA, Bashyam MD, Hidalgo M, et al.: Convergent structural alterations define SWItch/Sucrose NonFermentable (SWI/SNF) chromatin remodeler as a central tumor suppressive complex in pancreatic cancer.

    Proc Natl Acad Sci USA 2012, 109:E252-E259. OpenURL

  48. Job B, Bernheim A, Beau-Faller M, Camilleri-Broët S, Girard P, Hofman P, et al.: Genomic aberrations in lung adenocarcinoma in never smokers.

    PLoS One 2010, 5:e15145. OpenURL

  49. Abdel-Wahab O, Mullally A, Hedvat C, Garcia-Manero G, Patel J, Wadleigh M, et al.: Genetic characterization of TET1, TET2, and TET3 alterations in myeloid malignancies.

    Blood 2009, 114:144-147. OpenURL

  50. Al Sarakbi W, Sasi W, Jiang WG, Roberts T, Newbold RF, Mokbel K: The mRNA expression of SETD2 in human breast cancer: correlation with clinico-pathological parameters.

    BMC Cancer 2009, 9:290. OpenURL

  51. Bai J, Mei P-J, Liu H, Li C, Li W, Wu Y-P, et al.: BRG1 expression is increased in human glioma and controls glioma cell proliferation, migration and invasion in vitro.

    J Cancer Res Clin Oncol 2012, 138:991-998. OpenURL

  52. Saladi SV, Keenen B, Marathe HG, Qi H, Chin K-V, de la Serna IL: Modulation of extracellular matrix/adhesion molecule expression by BRG1 is associated with increased melanoma invasiveness.

    Mol Cancer 2010, 9:280. OpenURL

  53. Pasqualucci L, Trifonov V, Fabbri G, Ma J, Rossi D, Chiarenza A, et al.: Analysis of the coding genome of diffuse large B-cell lymphoma.

    Nat Genet 2011, 43:830-837. OpenURL

  54. Natarajan TG, Kallakury BV, Sheehan CE, Bartlett MB, Ganesan N, Preet A, et al.: Epigenetic regulator MLL2 shows altered expression in cancer cell lines and tumors from human breast and colon.

    Cancer Cell Int 2010, 10:13. OpenURL

  55. Kim MS, Chung NG, Kang MR, Yoo NJ, Lee SH: Genetic and expressional alterations of CHD genes in gastric and colorectal cancers.

    Histopathology 2011, 58:660-668. OpenURL

  56. Stephens PJ, Tarpey PS, Davies H, Loo PV, Greenman C, Wedge DC, et al.: The landscape of cancer genes and mutational processes in breast cancer.

    Nature 2012, 486:400-404. OpenURL

  57. Zhang Z, Yamashita H, Toyama T, Sugiura H, Ando Y, Mita K, et al.: NCOR1 mRNA is an independent prognostic factor for breast cancer.

    Cancer Lett 2006, 237:123-129. OpenURL

  58. Peifer M, Fernández-Cuesta L, Sos ML, George J, Seidel D, Kasper LH, et al.: Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer.

    Nat Genet 2012, 44:1104-1110. OpenURL

  59. Li Y, Yang H-X, Luo R-Z, Zhang Y, Li M, Wang X, et al.: High expression of p300 has an unfavorable impact on survival in resectable esophageal squamous cell carcinoma.

    Ann Thorac Surg 2011, 91:1531-1538. OpenURL

  60. Yokomizo C, Yamaguchi K, Itoh Y, Nishimura T, Umemura A, Minami M, et al.: High expression of p300 in HCC predicts shortened overall survival in association with enhanced epithelial mesenchymal transition of HCC cells.

    Cancer Lett 2011, 310:140-147. OpenURL

  61. Dokmanovic M, Clarke C, Marks PA: Histone deacetylase inhibitors: overview and perspectives.

    Mol Cancer Res 2007, 5:981-989. OpenURL

  62. Hodis E, Watson IR, Kryukov GV, Arold ST, Imielinski M, Theurillat J-P, et al.: A landscape of driver mutations in melanoma.

    Cell 2012, 150:251-263. OpenURL

  63. Manceau G, Letouzé E, Guichard C, Didelot A, Cazes A, Corté H, et al.: Recurrent inactivating mutations of ARID2 in non-small cell lung carcinoma.

    Int J Cancer 2012, 132:2217-2221. OpenURL

  64. Vainchenker W, Delhommeau F, Constantinescu SN, Bernard OA: New mutations and pathogenesis of myeloproliferative neoplasms.

    Blood 2011, 118:1723-1735. OpenURL

  65. Devillier R, Gelsi-Boyer V, Brecqueville M, Carbuccia N, Murati A, Vey N, et al.: Acute myeloid leukemia with myelodysplasia-related changes are characterized by a specific molecular pattern with high frequency of ASXL1 mutations.

    Am J Hematol 2012, 87:659-662. OpenURL

  66. Grasso CS, Wu Y-M, Robinson DR, Cao X, Dhanasekaran SM, Khan AP, et al.: The mutational landscape of lethal castrate resistant prostate cancer.

    Nature 2012, 487:239-243. OpenURL

  67. Hanlon K, Rudin CE, Harries LW: Investigating the targets of MIR-15a and MIR-16-1 in patients with Chronic Lymphocytic Leukemia (CLL).

    PLoS One 2009, 4:e7169. OpenURL

  68. Jiao Y, Killela PJ, Reitman ZJ, Rasheed BA, Heaphy CM, de Wilde RF, et al.: Frequent ATRX, CIC, FUBP1 and IDH1 mutations refine the classification of malignant gliomas.

    Oncotarget 2012, 3:709-722. OpenURL

  69. Xu J, Zhu W, Xu W, Yao W, Zhang B, Xu Y, et al.: Up-regulation of MBD1 promotes pancreatic cancer cell epithelial-mesenchymal transition and invasion by epigenetic down-regulation of E-cadherin.

    Curr Mol Med 2013, 13:387-400. OpenURL

  70. Fan L-H, Tang L-N, Yue L, Yang Y, Gao Z-L, Shen Z: BAP1 is a good prognostic factor in advanced non-small cell lung cancer.

    Clin Invest Med 2012, 35:E182-E189. OpenURL

  71. Quesada V, Conde L, Villamor N, Ordóñez GR, Jares P, Bassaganyas L, et al.: Exome sequencing identifies recurrent mutations of the splicing factor SF3B1 gene in chronic lymphocytic leukemia.

    Nat Genet 2012, 44:47-52. OpenURL

  72. Bandrés E, Malumbres R, Cubedo E, Honorato B, Zarate R, Labarga A, et al.: A gene signature of 8 genes could identify the risk of recurrence and progression in Dukes’ B colon cancer patients.

    Oncol Rep 2007, 17:1089-1094. OpenURL

  73. Skawran B, Steinemann D, Weigmann A, Flemming P, Becker T, Flik J, et al.: Gene expression profiling in hepatocellular carcinoma: upregulation of genes in amplified chromosome regions.

    Mod Pathol 2008, 21:505-516. OpenURL

  74. Buganim Y, Goldstein I, Lipson D, Milyavsky M, Polak-Charcon S, Mardoukh C, et al.: A novel translocation breakpoint within the BPTF gene is associated with a pre-malignant phenotype.

    PLoS One 2010, 5:e9657. OpenURL

  75. Deshpande AM, Akunowicz JD, Reveles XT, Patel BB, Saria EA, Gorlick RG, et al.: PHC3, a component of the hPRC-H complex, associates with E2F6 during G0 and is lost in osteosarcoma tumors.

    Oncogene 2007, 26:1714-1722. OpenURL

  76. Moloney FJ, Lyons JG, Bock VL, Huang XX, Bugeja MJ, Halliday GM: Hotspot mutation of Brahma in non-melanoma skin cancer.

    J Invest Dermatol 2009, 129:1012-1015. OpenURL

  77. Doménech E, Gómez-López G, Gzlez-Peña D, López M, Herreros B, Menezes J, et al.: New mutations in chronic lymphocytic leukemia identified by target enrichment and deep sequencing.

    PLoS One 2012, 7:e38158. OpenURL

  78. Reisman DN, Sciarrotta J, Wang W, Funkhouser WK, Weissman BE: Loss of BRG1/BRM in human lung cancer cell lines and primary lung cancers: correlation with poor prognosis.

    Cancer Res 2003, 63:560-566. OpenURL

  79. Yamamichi N, Inada K, Ichinose M, Yamamichi-Nishina M, Mizutani T, Watanabe H, et al.: Frequent loss of Brm expression in gastric cancer correlates with histologic features and differentiation state.

    Cancer Res 2007, 67:10727-10735. OpenURL

  80. Hélias C, Struski S, Gervais C, Leymarie V, Mauvieux L, Herbrecht R, et al.: Polycythemia vera transforming to acute myeloid leukemia and complex abnormalities including 9p homogeneously staining region with amplification of MLLT3, JMJD2C, JAK2, and SMARCA2.

    Cancer Genet Cytogenet 2008, 180:51-55. OpenURL

  81. Ceol CJ, Houvras Y, Jane-Valbuena J, Bilodeau S, Orlando DA, Battisti V, et al.: The histone methyltransferase SETDB1 is recurrently amplified in melanoma and accelerates its onset.

    Nature 2011, 471:513-517. OpenURL

  82. Futreal PA, Coin L, Marshall M, Down T, Hubbard T, Wooster R, et al.: A census of human cancer genes.

    Nat Rev Cancer 2004, 4:177-183. OpenURL

  83. Gonzalez-Perez A, Perez-Llamas C, Deu-Pons J, Tamborero D, Schroeder MP, Jene-Sanz A, et al.: IntOGen-mutations identifies cancer drivers across tumor types.

    Nat Meth 2013, 5:5. OpenURL

  84. McLendon R, Friedman A, Bigner D, Meir EGV, Brat DJ, Mastrogianakis GM, et al.: Comprehensive genomic characterization defines human glioblastoma genes and core pathways.

    Nature 2008, 455:1061-1068. OpenURL

  85. Hudson TJ, Anderson W, Aretz A, Barker AD, Bell C, Bernabé RR, et al.: International network of cancer genome projects.

    Nature 2010, 464:993-998. OpenURL

  86. Ben-Porath I, Thomson MW, Carey VJ, Ge R, Bell GW, Regev A, et al.: An embryonic stem cell-like gene expression signature in poorly differentiated aggressive human tumors.

    Nat Genet 2008, 40:499-507. OpenURL

  87. Gundem G, Perez-Llamas C, Jene-Sanz A, Kedzierska A, Islam A, Deu-Pons J, et al.: IntOGen: Integration and data-mining of multidimensional oncogenomic data.

    Nat Meth 2010, 7:92-93. OpenURL

  88. Parsons DW, Jones S, Zhang X, Lin JC-H, Leary RJ, Angenendt P, et al.: An integrated genomic analysis of human glioblastoma multiforme.

    Science 2008, 321:1807-1812. OpenURL

  89. Rausch T, Jones DTW, Zapatka M, Stütz AM, Zichner T, Weischenfeldt J, et al.: Genome sequencing of pediatric medulloblastoma links catastrophic DNA rearrangements with TP53 mutations.

    Cell 2012, 148:59-71. OpenURL

  90. Schwartzentruber J, Korshunov A, Liu X-Y, Jones DTW, Pfaff E, Jacob K, et al.: Driver mutations in histone H3.3 and chromatin remodelling genes in paediatric glioblastoma.

    Nature 2012, 482:226-231. OpenURL

  91. Wood LD, Parsons DW, Jones S, Lin J, Sjöblom T, Leary RJ, et al.: The genomic landscapes of human breast and colorectal cancers.

    Science 2007, 318:1108-1113. OpenURL

  92. Shah SP, Roth A, Goya R, Oloumi A, Ha G, Zhao Y, et al.: The clonal and mutational evolution spectrum of primary triple-negative breast cancers.

    Nature 2012, 486:395-399. OpenURL

  93. The Cancer Genome Atlas Network: Comprehensive molecular portraits of human breast tumours.

    Nature 2012, 490:61-70. OpenURL

  94. Banerji S, Cibulskis K, Rangel-Escareno C, Brown KK, Carter SL, Frederick AM, et al.: Sequence analysis of mutations and translocations across breast cancer subtypes.

    Nature 2012, 486:405-409. OpenURL

  95. Stransky N, Egloff AM, Tward AD, Kostic AD, Cibulskis K, Sivachenko A, et al.: The mutational landscape of head and neck squamous cell carcinoma.

    Science 2011, 333:1157-1160. OpenURL

  96. Puente XS, Pinyol M, Quesada V, Conde L, Ordóñez GR, Villamor N, et al.: Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia.

    Nature 2011, 475:101-105. OpenURL

  97. Wang L, Lawrence MS, Wan Y, Stojanov P, Sougnez C, Stevenson K, et al.: SF3B1 and other novel cancer genes in chronic lymphocytic leukemia.

    New Engl J Med 2011, 365:2497-2506. OpenURL

  98. The Cancer Genome Atlas Research Network: Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia.

    New Engl J Med 2013, 368:2059-2074. OpenURL

  99. The Cancer Genome Atlas Research Network: Comprehensive molecular characterization of clear cell renal cell carcinoma.

    Nature 2013, 499:43-49. OpenURL

  100. Guichard C, Amaddeo G, Imbeaud S, Ladeiro Y, Pelletier L, Maad IB, et al.: Integrated analysis of somatic mutations and focal copy-number changes identifies key genes and pathways in hepatocellular carcinoma.

    Nat Genet 2012, 44:694-698. OpenURL

  101. Ding L, Getz G, Wheeler DA, Mardis ER, McLellan MD, Cibulskis K, et al.: Somatic mutations affect key pathways in lung adenocarcinoma.

    Nature 2008, 455:1069-1075. OpenURL

  102. The Cancer Genome Atlas Research Network: Comprehensive genomic characterization of squamous cell lung cancers.

    Nature 2012, 489:519-525. OpenURL

  103. Rudin CM, Durinck S, Stawiski EW, Poirier JT, Modrusan Z, Shames DS, et al.: Comprehensive genomic analysis identifies SOX2 as a frequently amplified gene in small-cell lung cancer.

    Nat Genet 2012, 44:1111-1116. OpenURL

  104. The Cancer Genome Atlas Research Network: Integrated genomic analyses of ovarian carcinoma.

    Nature 2011, 474:609-615. OpenURL

  105. Jones S, Zhang X, Parsons DW, Lin JC-H, Leary RJ, Angenendt P, et al.: Core signaling pathways in human pancreatic cancers revealed by global genomic analyses.

    Science 2008, 321:1801-1806. OpenURL

  106. ICGC data portal [http://dcc.icgc.org webcite]

  107. Wang K, Kan J, Yuen ST, Shi ST, Chu KM, Law S, et al.: Exome sequencing identifies frequent mutation of ARID1A in molecular subtypes of gastric cancer.

    Nat Genet 2011, 43:1219-1223. OpenURL

  108. Wu G, Feng X, Stein L: A human functional protein interaction network and its application to cancer data analysis.

    Genome Biol 2010, 11:R53. OpenURL

  109. Reva B, Antipin Y, Sander C: Predicting the functional impact of protein mutations: application to cancer genomics.

    Nucl Acids Res 2011, 39:e118-e118. OpenURL

  110. Gundem G, Lopez-Bigas N: Sample level enrichment analysis unravels shared stress phenotypes among multiple cancer types.

    Genome Med 2012, 4:28. OpenURL

  111. Schinzel AC, Hahn WC: Oncogenic transformation and experimental models of human cancer.

    Front Biosci 2008, 13:71-84. OpenURL

  112. Integrative Onco Genomics: mutations [http://www.intogen.org/mutations webcite]

  113. Gitools IntOgen datasets [http://www.gitools.org/datasets#IntOGen webcite]

  114. The UniProt Consortium: Reorganizing the protein space at the Universal Protein Resource (UniProt).

    Nucleic Acids Res 2011, 40:D71-D75. OpenURL

  115. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al.: Cytoscape: a software environment for integrated models of biomolecular interaction networks.

    Genome Res 2003, 13:2498-2504. OpenURL

  116. Perez-Llamas C, Lopez-Bigas N: Gitools: analysis and visualisation of genomic data using interactive heat-maps.

    PLoS ONE 2011, 6:e19541. OpenURL

  117. Lopez-Bigas N, De S, Teichmann SA: Functional protein divergence in the evolution of Homo sapiens.

    Genome Biol 2008, 9:R33. OpenURL

  118. Mann HB, Whitney DR: On a test of whether one of two random variables is stochastically larger than the other.

    Ann Math Statist 1947, 18:50-60. OpenURL

  119. Gitools [http://www.gitools.org webcite]

  120. Broad-novartis cancer cell line encyclopedia [http://www.broadinstitute.org/ccle webcite]

  121. The ENCODE Project Consortium: Identification and analysis of functional elements in 1% of the human genome by the ENCODE pilot project.

    Nature 2007, 447:799-816. OpenURL

  122. Lister R, Pelizzola M, Dowen RH, Hawkins RD, Hon G, Tonti-Filippini J, et al.: Human DNA methylomes at base resolution show widespread epigenomic differences.

    Nature 2009, 462:315-322. OpenURL

  123. Barski A, Cuddapah S, Cui K, Roh T-Y, Schones DE, Wang Z, et al.: High-resolution profiling of histone methylations in the human genome.

    Cell 2007, 129:823-837. OpenURL

  124. Viré E, Brenner C, Deplus R, Blanchon L, Fraga M, Didelot C, et al.: The Polycomb group protein EZH2 directly controls DNA methylation.

    Nature 2005, 439:871-874. OpenURL

  125. Hansen RS, Thomas S, Sandstrom R, Canfield TK, Thurman RE, Weaver M, et al.: Sequencing newly replicated DNA reveals widespread plasticity in human replication timing.

    Proc Natl Acad Sci 2010, 107:139-144. OpenURL

  126. Langmead B, Trapnell C, Pop M, Salzberg SL: Ultrafast and memory-efficient alignment of short DNA sequences to the human genome.

    Genome Biol 2009, 10:R25. OpenURL

  127. Zhang Y, Liu T, Meyer C, Eeckhoute J, Johnson D, Bernstein B, et al.: Model-based Analysis of ChIP-Seq (MACS).

    Genome Biol 2008, 9:R137. OpenURL

  128. Zang C, Schones DE, Zeng C, Cui K, Zhao K, Peng W: A clustering approach for identification of enriched domains from histone modification ChIP-Seq data.

    Bioinformatics 2009, 25:1952-1958. OpenURL

  129. Quinlan AR, Hall IM: BEDTools: a flexible suite of utilities for comparing genomic features.

    Bioinformatics 2010, 26:841-842. OpenURL

  130. Ji X, Li W, Song J, Wei L, Liu XS: CEAS: cis-regulatory element annotation system.

    Nucl Acids Res 2006, 34:W551-W554. OpenURL

  131. Kanehisa M, Goto S, Sato Y, Furumichi M, Tanabe M: KEGG for integration and interpretation of large-scale molecular data sets.

    Nucleic Acids Res 2012, 40:D109-D114. OpenURL

  132. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al.: Gene Ontology: tool for the unification of biology.

    Nat Genet 2000, 25:25-29. OpenURL