The NADPH oxidase NOX4 regulates redox and metabolic homeostasis preventing HCC progression.

BACKGROUND AND AIMS: The NADPH oxidase NOX4 plays a tumor-suppressor function in HCC. Silencing NOX4 confers higher proliferative and migratory capacity to HCC cells and increases their in vivo tumorigenic potential in xenografts in mice. NOX4 gene deletions are frequent in HCC, correlating with higher tumor grade and worse recurrence-free and overall survival rates. However, despite the accumulating evidence of a protective regulatory role in HCC, the cellular processes governed by NOX4 are not yet understood. Accordingly, the aim of this work was to better understand the molecular mechanisms regulated by NOX4 in HCC in order to explain its tumor-suppressor action. APPROACH AND RESULTS: Experimental models: cell-based loss or gain of NOX4 function experiments, in vivo hepatocarcinogenesis induced by diethylnitrosamine in Nox4 -deficient mice, and analyses in human HCC samples. Methods include cellular and molecular biology analyses, proteomics, transcriptomics, and metabolomics, as well as histological and immunohistochemical analyses in tissues. Results identified MYC as being negatively regulated by NOX4. MYC mediated mitochondrial dynamics and a transcriptional program leading to increased oxidative metabolism, enhanced use of both glucose and fatty acids, and an overall higher energetic capacity and ATP level. NOX4 deletion induced a redox imbalance that augmented nuclear factor erythroid 2-related factor 2 (Nrf2) activity and was responsible for MYC up-regulation. CONCLUSIONS: Loss of NOX4 in HCC tumor cells induces metabolic reprogramming in a Nrf2/MYC-dependent manner to promote HCC progression.

Clathrin switches transforming growth factor-ß role to pro-tumorigenic in liver cancer.

BACKGROUND & AIMS: Upon ligand binding, tyrosine kinase receptors, such as epidermal growth factor receptor (EGFR), are recruited into clathrin-coated pits for internalization by endocytosis, which is relevant for signalling and/or receptor degradation. In liver cells, transforming growth factor-ß (TGF-ß) induces both pro- and anti-apoptotic signals; the latter are mediated by the EGFR pathway. Since EGFR mainly traffics via clathrin-coated vesicles, we aimed to analyse the potential role of clathrin in TGF-ß-induced signalling in liver cells and its relevance in liver cancer. METHODS: Real-Time PCR and immunohistochemistry were used to analyse clathrin heavy-chain expression in human (CLTC) and mice (Cltc) liver tumours. Transient knockdown (siRNA) or overexpression of CLTC were used to analyse its role on TGF-ß and EGFR signalling in vitro. Bioinformatic analysis was used to determine the effect of CLTC and TGFB1 expression on prognosis and overall survival in patients with hepatocellular carcinoma (HCC). RESULTS: Clathrin expression increased during liver tumorigenesis in humans and mice. CLTC knockdown cells responded to TGF-ß phosphorylating SMADs (canonical signalling) but showed impairment in the anti-apoptotic signals (EGFR transactivation). Experiments of loss or gain of function in HCC cells reveal an essential role for clathrin in inhibiting TGF-ß-induced apoptosis and upregulation of its pro-apoptotic target NOX4. Autocrine TGF-ß signalling in invasive HCC cells upregulates CLTC expression, switching its role to pro-tumorigenic. A positive correlation between TGFB1 and CLTC was found in HCC cells and patients. Patients expressing high levels of TGFB1 and CLTC had a worse prognosis and lower overall survival. CONCLUSIONS: This work describes a novel role for clathrin in liver tumorigenesis, favouring non-canonical pro-tumorigenic TGF-ß pathways. CLTC expression in human HCC samples could help select patients that would benefit from TGF-ß-targeted therapy. LAY SUMMARY: Clathrin heavy-chain expression increases during liver tumorigenesis in humans (CLTC) and mice (Cltc), altering the cellular response to TGF-ß in favour of anti-apoptotic/pro-tumorigenic signals. A positive correlation between TGFB1 and CLTC was found in HCC cells and patients. Patients expressing high levels of TGFB1 and CLTC had a worse prognosis and lower overall survival. CLTC expression in HCC human samples could help select patients that would benefit from therapies targeting TGF-ß.

Colon-specific eQTL analysis to inform on functional SNPs.

BACKGROUND: Genome-wide association studies on colorectal cancer have identified more than 60 susceptibility loci, but for most of them there is no clear knowledge of functionality or the underlying gene responsible for the risk modification. Expression quantitative trail loci (eQTL) may provide functional information for such single nucleotide polymorphisms (SNPs). METHODS: We have performed detailed eQTL analysis specific for colon tissue on a series of 97 colon tumours, their paired adjacent normal mucosa and 47 colon mucosa samples donated by healthy individuals. R package MatrixEQTL was used to search for genome-wide cis-eQTL and trans-eQTL fitting linear models adjusted for age, gender and tissue type to rank transformed expression data. RESULTS: The cis-eQTL analyses has revealed 29,073 SNP-gene associations with permutation-adjusted P-values < 0.01. These correspond to 363 unique genes. The trans-eQTL analysis identified 10,665 significant SNP-gene associations, most of them in the same chromosome, further than 1 Mb of the gene. We provide a web tool to search for specific SNPs or genes. The tool calculates Pearson or Spearman correlation, and allows to select tissue type for analysis. Data and plots can be exported. CONCLUSIONS: This resource should be useful to prioritise SNPs for further functional studies and to identify relevant genes behind identified loci.

AKT1low quiescent cancer cells persist after neoadjuvant chemotherapy in triple negative breast cancer.

BACKGROUND: Absence of pathologic complete response (pCR) to neoadjuvant chemotherapy (NACT) correlates with poor long-term survival in patients with triple negative breast cancer (TNBC). These incomplete treatment responses are likely determined by mechanisms that enable cancer cells to resist being killed. However, the detailed characterization of a drug-resistant cancer cell state in residual TNBC tissue after NACT has remained elusive. AKT1low quiescent cancer cells (QCCs) are a quiescent, epigenetically plastic, and chemotherapy-resistant subpopulation initially identified in experimental cancer models. Here, we asked whether QCCs exist in primary tumors from patients with TNBC and persist after treatment with NACT. METHODS: We obtained pre-treatment biopsy, post-treatment mastectomy, and metastatic specimens from a retrospective cohort of TNBC patients treated with NACT at Massachusetts General Hospital (n = 25). Using quantitative automated immunofluorescence microscopy, QCCs were identified as AKTlow/H3K9me2low/HES1high cancer cells using prespecified immunofluorescence intensity thresholds. QCCs were represented in 2D and 3D digital tumor maps and QCC percentage (QCC-P) and QCC cluster index (QCC-CI) were determined for each sample. RESULTS: We showed that QCCs exist as non-random and heterogeneously distributed clusters within primary breast tumors. In addition, these QCC clusters persist after treatment with multi-agent, multi-cycle, neoadjuvant chemotherapy in both residual primary tumors and nodal and distant metastases in patients with triple negative breast cancer. CONCLUSIONS: These first-in-human data potentially qualify AKT1low quiescent cancer cells as a non-genetic cell state that persists after neoadjuvant chemotherapy in triple negative breast cancer patients and warrants further study.

Comprehensive analysis of copy number aberrations in microsatellite stable colon cancer in view of stromal component.

BACKGROUND: Somatic copy number aberrations (CNAs) are common acquired changes in cancer cells having an important role in the progression of colon cancer (colorectal cancer, CRC). This study aimed to perform a characterisation of CNA and their impact in gene expression. METHODS: Copy number aberrations were inferred from SNP array data in a series of 99 CRC. Copy number aberration events were calculated and used to assess the association between copy number dosage, clinical and molecular characteristics of the tumours, and gene expression changes. All analyses were adjusted for the quantity of stroma in each sample, which was inferred from gene expression data. RESULTS: High heterogeneity among samples was observed; the proportion of altered genome ranged between 0.04 and 26.6%. Recurrent CNA regions with gains were frequent in chromosomes 7p, 8q, 13q, and 20, whereas 8p, 17p, and 18 cumulated losses. A significant positive correlation was observed between the number of somatic mutations and total CNA (Spearman's r=0.42, P=0.006). Approximately 37% of genes located in CNA regions changed their level of expression and the average partial correlation (adjusted for stromal content) with copy number was 0.54 (interquartile range 0.20 to 0.81). Altered genes showed enrichment in pathways relevant for CRC. Tumours classified as CMS2 and CMS4 by the consensus molecular subtyping showed higher frequency of CNA. Losses of one small region in 1p36.33, with gene CDK11B, were associated with poor prognosis. More than 66% of the recurrent CNA were validated in the The Cancer Genome Atlas (TCGA) data when analysed with the same procedure. Furthermore, 79% of the genes with altered expression in our data were validated in the TCGA. CONCLUSIONS: Although CNA are frequent events in microsatellite stable CRC, few focal recurrent regions were found. These aberrations have strong effects on gene expression and contribute to deregulate relevant cancer pathways. Owing to the diploid nature of stromal cells, it is important to consider the purity of tumour samples to accurately calculate CNA events in CRC.

Identification of candidate susceptibility genes for colorectal cancer through eQTL analysis.

In this study, we aim to identify the genes responsible for colorectal cancer risk behind the loci identified in genome-wide association studies (GWAS). These genes may be candidate targets for developing new strategies for prevention or therapy. We analyzed the association of genotypes for 26 GWAS single nucleotide polymorphisms (SNPs) with the expression of genes within a 2 Mb region (cis-eQTLs). Affymetrix Human Genome U219 expression arrays were used to assess gene expression in two series of samples, one of healthy colonic mucosa (n = 47) and other of normal mucosa adjacent to colon cancer (n = 97, total 144). Paired tumor tissues (n = 97) were also analyzed but did not provide additional findings. Partial Pearson correlation (r), adjusted for sample type, was used for the analysis. We have found Bonferroni-significant cis-eQTLs in three loci: rs3802842 in 11q23.1 associated to C11orf53, COLCA1 (C11orf92) and COLCA2 (C11orf93; r = 0.60); rs7136702 in 12q13.12 associated to DIP2B (r = 0.63) and rs5934683 in Xp22.3 associated to SHROOM2 and GPR143 (r = 0.47). For loci in chromosomes 11 and 12, we have found other SNPs in linkage disequilibrium that are more strongly associated with the expression of the identified genes and are better functional candidates: rs7130173 for 11q23.1 (r = 0.66) and rs61927768 for 12q13.12 (r = 0.86). These SNPs are located in DNA regions that may harbor enhancers or transcription factor binding sites. The analysis of trans-eQTLs has identified additional genes in these loci that may have common regulatory mechanisms as shown by the analysis of protein-protein interaction networks.

Aberrant gene expression in mucosa adjacent to tumor reveals a molecular crosstalk in colon cancer.

BACKGROUND: A colorectal tumor is not an isolated entity growing in a restricted location of the body. The patient's gut environment constitutes the framework where the tumor evolves and this relationship promotes and includes a complex and tight correlation of the tumor with inflammation, blood vessels formation, nutrition, and gut microbiome composition. The tumor influence in the environment could both promote an anti-tumor or a pro-tumor response. METHODS: A set of 98 paired adjacent mucosa and tumor tissues from colorectal cancer (CRC) patients and 50 colon mucosa from healthy donors (246 samples in total) were included in this work. RNA extracted from each sample was hybridized in Affymetrix chips Human Genome U219. Functional relationships between genes were inferred by means of systems biology using both transcriptional regulation networks (ARACNe algorithm) and protein-protein interaction networks (BIANA software). RESULTS: Here we report a transcriptomic analysis revealing a number of genes activated in adjacent mucosa from CRC patients, not activated in mucosa from healthy donors. A functional analysis of these genes suggested that this active reaction of the adjacent mucosa was related to the presence of the tumor. Transcriptional and protein-interaction networks were used to further elucidate this response of normal gut in front of the tumor, revealing a crosstalk between proteins secreted by the tumor and receptors activated in the adjacent colon tissue; and vice versa. Remarkably, Slit family of proteins activated ROBO receptors in tumor whereas tumor-secreted proteins transduced a cellular signal finally activating AP-1 in adjacent tissue. CONCLUSIONS: The systems-level approach provides new insights into the micro-ecology of colorectal tumorogenesis. Disrupting this intricate molecular network of cell-cell communication and pro-inflammatory microenvironment could be a therapeutic target in CRC patients.

Large differences in global transcriptional regulatory programs of normal and tumor colon cells.

BACKGROUND: Dysregulation of transcriptional programs leads to cell malfunctioning and can have an impact in cancer development. Our study aims to characterize global differences between transcriptional regulatory programs of normal and tumor cells of the colon. METHODS: Affymetrix Human Genome U219 expression arrays were used to assess gene expression in 100 samples of colon tumor and their paired adjacent normal mucosa. Transcriptional networks were reconstructed using ARACNe algorithm using 1,000 bootstrap replicates consolidated into a consensus network. Networks were compared regarding topology parameters and identified well-connected clusters. Functional enrichment was performed with SIGORA method. ENCODE ChIP-Seq data curated in the hmChIP database was used for in silico validation of the most prominent transcription factors. RESULTS: The normal network contained 1,177 transcription factors, 5,466 target genes and 61,226 transcriptional interactions. A large loss of transcriptional interactions in the tumor network was observed (11,585; 81% reduction), which also contained fewer transcription factors (621; 47% reduction) and target genes (2,190; 60% reduction) than the normal network. Gene silencing was not a main determinant of this loss of regulatory activity, since the average gene expression was essentially conserved. Also, 91 transcription factors increased their connectivity in the tumor network. These genes revealed a tumor-specific emergent transcriptional regulatory program with significant functional enrichment related to colorectal cancer pathway. In addition, the analysis of clusters again identified subnetworks in the tumors enriched for cancer related pathways (immune response, Wnt signaling, DNA replication, cell adherence, apoptosis, DNA repair, among others). Also multiple metabolism pathways show differential clustering between the tumor and normal network. CONCLUSIONS: These findings will allow a better understanding of the transcriptional regulatory programs altered in colon cancer and could be an invaluable methodology to identify potential hubs with a relevant role in the field of cancer diagnosis, prognosis and therapy.

Transcriptional analysis of landmark molecular pathways in lung adenocarcinoma results in a clinically relevant classification with potential therapeutic implications.

Lung adenocarcinoma (LUAD) is a molecularly heterogeneous disease. In addition to genomic alterations, cancer transcriptional profiling can be helpful to tailor cancer treatment and to estimate each patient's outcome. Transcriptional activity levels of 50 molecular pathways were inferred in 4573 LUAD patients using Gene Set Variation Analysis (GSVA) method. Seven LUAD subtypes were defined and independently validated based on the combined behavior of the studied pathways: AD (adenocarcinoma subtype) 1-7. AD1, AD4, and AD5 subtypes were associated with better overall survival. AD1 and AD4 subtypes were enriched in epidermal growth factor receptor (EGFR) mutations, whereas AD2 and AD6 showed higher tumor protein p53 (TP53) alteration frequencies. AD2 and AD6 subtypes correlated with higher genome instability, proliferation-related pathway expression, and specific sensitivity to chemotherapy, based on data from LUAD cell lines. LUAD subtypes were able to predict immunotherapy response in addition to CD274 (PD-L1) gene expression and tumor mutational burden (TMB). AD2 and AD4 subtypes were associated with potential resistance and response to immunotherapy, respectively. Thus, analysis of transcriptomic data could improve patient stratification beyond genomics and single biomarkers (i.e., PD-L1 and TMB) and may lay the foundation for more personalized treatment avenues, especially in driver-negative LUAD.

Identification of a Twelve-microRNA Signature with Prognostic Value in Stage II Microsatellite Stable Colon Cancer.

We aimed to identify and validate a set of miRNAs that could serve as a prognostic signature useful to determine the recurrence risk for patients with COAD. Small RNAs from tumors of 100 stage II, untreated, MSS colon cancer patients were sequenced for the discovery step. For this purpose, we built an miRNA score using an elastic net Cox regression model based on the disease-free survival status. Patients were grouped into high or low recurrence risk categories based on the median value of the score. We then validated these results in an independent sample of stage II microsatellite stable tumor tissues, with a hazard ratio of 3.24, (CI95% = 1.05-10.0) and a 10-year area under the receiver operating characteristic curve of 0.67. Functional analysis of the miRNAs present in the signature identified key pathways in cancer progression. In conclusion, the proposed signature of 12 miRNAs can contribute to improving the prediction of disease relapse in patients with stage II MSS colorectal cancer, and might be useful in deciding which patients may benefit from adjuvant chemotherapy.

COLONOMICS - integrative omics data of one hundred paired normal-tumoral samples from colon cancer patients.

Colonomics is a multi-omics dataset that includes 250 samples: 50 samples from healthy colon mucosa donors and 100 paired samples from colon cancer patients (tumor/adjacent). From these samples, Colonomics project includes data from genotyping, DNA methylation, gene expression, whole exome sequencing and micro-RNAs (miRNAs) expression. It also includes data from copy number variation (CNV) from tumoral samples. In addition, clinical data from all these samples is available. The aims of the project were to explore and integrate these datasets to describe colon cancer at molecular level and to compare normal and tumoral tissues. Also, to improve screening by finding biomarkers for the diagnosis and prognosis of colon cancer. This project has its own website including four browsers allowing users to explore Colonomics datasets. Since generated data could be reuse for the scientific community for exploratory or validation purposes, here we describe omics datasets included in the Colonomics project as well as results from multi-omics layers integration.

Gene Expression Profiling as a Potential Tool for Precision Oncology in Non-Small Cell Lung Cancer.

Recent technological advances and the application of high-throughput mutation and transcriptome analyses have improved our understanding of cancer diseases, including non-small cell lung cancer. For instance, genomic profiling has allowed the identification of mutational events which can be treated with specific agents. However, detection of DNA alterations does not fully recapitulate the complexity of the disease and it does not allow selection of patients that benefit from chemo- or immunotherapy. In this context, transcriptional profiling has emerged as a promising tool for patient stratification and treatment guidance. For instance, transcriptional profiling has proven to be especially useful in the context of acquired resistance to targeted therapies and patients lacking targetable genomic alterations. Moreover, the comprehensive characterization of the expression level of the different pathways and genes involved in tumor progression is likely to better predict clinical benefit from different treatments than single biomarkers such as PD-L1 or tumor mutational burden in the case of immunotherapy. However, intrinsic technical and analytical limitations have hindered the use of these expression signatures in the clinical setting. In this review, we will focus on the data reported on molecular classification of non-small cell lung cancer and discuss the potential of transcriptional profiling as a predictor of survival and as a patient stratification tool to further personalize treatments.

Breast cancer dormancy is associated with a 4NG1 state and not senescence.

Reactivation of dormant cancer cells can lead to cancer relapse, metastasis, and patient death. Dormancy is a nonproliferative state and is linked to late relapse and death. No targeted therapy is currently available to eliminate dormant cells, highlighting the need for a deeper understanding and reliable models. Here, we thoroughly characterize the dormant D2.OR and ZR-75-1, and proliferative D2A1 breast cancer cell line models in vivo and/or in vitro, and assess if there is overlap between a dormant and a senescent phenotype. We show that D2.OR but not D2A1 cells become dormant in the liver of an immunocompetent model. In vitro, we show that D2.OR and ZR-75-1 cells in response to a 3D environment or serum-free conditions are growth-arrested in G1, of which a subpopulation resides in a 4NG1 state. The dormancy state is reversible and not associated with a senescence phenotype. This will aid future research on breast cancer dormancy.

Integrative transcriptome analysis of malignant pleural mesothelioma reveals a clinically relevant immune-based classification.

BACKGROUND: Malignant pleural mesothelioma (MPM) is a rare and aggressive neoplasia affecting the lung mesothelium. Immune checkpoint inhibitors (ICI) in MPM have not been extremely successful, likely due to poor identification of suitable candidate patients for the therapy. We aimed to identify cellular immune fractions associated with clinical outcome and classify patients with MPM based on their immune contexture. For each defined group, we sought for molecular specificities that could help further define our MPM classification at the genomic and transcriptomic level, as well as identify differential therapeutic strategies based on transcriptional signatures predictive of drug response. METHODS: The abundance of 20 immune cell fractions in 516 MPM samples from 7 gene expression datasets was inferred using gene set variation analysis. Identification of clinically relevant fractions was performed with Cox proportional-hazards models adjusted for age, stage, sex, and tumor histology. Immune-based groups were defined based on the identified fractions. RESULTS: T-helper 2 (TH2) and cytotoxic T (TC) cells were found to be consistently associated with overall survival. Three immune clusters (IG) were subsequently defined based on TH2 and TC immune infiltration levels: IG1 (54.5%) was characterized by high TH2 and low TC levels, IG2 (37%) had either low or high levels of both fractions, and IG3 (8.5%) was defined by low TH2 and high TC levels. IG1 and IG3 groups were associated with worse and better overall survival, respectively. While no differential genomic alterations were identified among immune groups, at the transcriptional level, IG1 samples showed upregulation of proliferation signatures, while IG3 samples presented upregulation of immune and inflammation-related pathways. Finally, the integration of gene expression with functional signatures of drug response showed that IG3 patients might be more likely to respond to ICI. CONCLUSIONS: This study identifies a novel immune-based signature with potential clinical relevance based on TH2 and TC levels, unveiling a fraction of patients with MPM with better prognosis and who might benefit from immune-based therapies. Molecular specificities of the different groups might be used to tailor specific potential therapies in the future.

DNA methylation events in transcription factors and gene expression changes in colon cancer.

Aim: Gain insight about the role of DNA methylation in the malignant growth of colon cancer. Patients & methods: Methylation and gene expression from 90 adjacent-tumor paired tissues and 48 healthy tissues were analyzed. Tumor genes whose change in expression was explained by changes in methylation were identified using linear models adjusted for tumor stromal content. Results: No differences in methylation were found between adjacent and healthy tissues, but clear differences were found between adjacent and tumor samples. We identified hypermethylated CpG islands located in promoter regions that drive differential gene expression of transcription factors and their target genes. Conclusion: Changes in methylation of a few genes provoke important changes in gene expression, by expanding the signal through transcription activation/repression.

Genetic and genomic analysis modeling of germline c-MYC overexpression and cancer susceptibility.

BACKGROUND: Germline genetic variation is associated with the differential expression of many human genes. The phenotypic effects of this type of variation may be important when considering susceptibility to common genetic diseases. Three regions at 8q24 have recently been identified to independently confer risk of prostate cancer. Variation at 8q24 has also recently been associated with risk of breast and colorectal cancer. However, none of the risk variants map at or relatively close to known genes, with c-MYC mapping a few hundred kilobases distally. RESULTS: This study identifies cis-regulators of germline c-MYC expression in immortalized lymphocytes of HapMap individuals. Quantitative analysis of c-MYC expression in normal prostate tissues suggests an association between overexpression and variants in Region 1 of prostate cancer risk. Somatic c-MYC overexpression correlates with prostate cancer progression and more aggressive tumor forms, which was also a pathological variable associated with Region 1. Expression profiling analysis and modeling of transcriptional regulatory networks predicts a functional association between MYC and the prostate tumor suppressor KLF6. Analysis of MYC/Myc-driven cell transformation and tumorigenesis substantiates a model in which MYC overexpression promotes transformation by down-regulating KLF6. In this model, a feedback loop through E-cadherin down-regulation causes further transactivation of c-MYC. CONCLUSION: This study proposes that variation at putative 8q24 cis-regulator(s) of transcription can significantly alter germline c-MYC expression levels and, thus, contribute to prostate cancer susceptibility by down-regulating the prostate tumor suppressor KLF6 gene.

Genetic interactions: the missing links for a better understanding of cancer susceptibility, progression and treatment.

It is increasingly clear that complex networks of relationships between genes and/or proteins govern neoplastic processes. Our understanding of these networks is expanded by the use of functional genomic and proteomic approaches in addition to computational modeling. Concurrently, whole-genome association scans and mutational screens of cancer genomes identify novel cancer genes. Together, these analyses have vastly increased our knowledge of cancer, in terms of both "part lists" and their functional associations. However, genetic interactions have hitherto only been studied in depth in model organisms and remain largely unknown for human systems. Here, we discuss the importance and potential benefits of identifying genetic interactions at the human genome level for creating a better understanding of cancer susceptibility and progression and developing novel effective anticancer therapies. We examine gene expression profiles in the presence and absence of co-amplification of the 8q24 and 20q13 chromosomal regions in breast tumors to illustrate the molecular consequences and complexity of genetic interactions and their role in tumorigenesis. Finally, we highlight current strategies for targeting tumor dependencies and outline potential matrix screening designs for uncovering molecular vulnerabilities in cancer cells.

SNPassoc: an R package to perform whole genome association studies.

UNLABELLED: The popularization of large-scale genotyping projects has led to the widespread adoption of genetic association studies as the tool of choice in the search for single nucleotide polymorphisms (SNPs) underlying susceptibility to complex diseases. Although the analysis of individual SNPs is a relatively trivial task, when the number is large and multiple genetic models need to be explored it becomes necessary a tool to automate the analyses. In order to address this issue, we developed SNPassoc, an R package to carry out most common analyses in whole genome association studies. These analyses include descriptive statistics and exploratory analysis of missing values, calculation of Hardy-Weinberg equilibrium, analysis of association based on generalized linear models (either for quantitative or binary traits), and analysis of multiple SNPs (haplotype and epistasis analysis). AVAILABILITY: Package SNPassoc is available at CRAN from http://cran.r-project.org. SUPPLEMENTARY INFORMATION: A tutorial is available on Bioinformatics online and in http://davinci.crg.es/estivill_lab/snpassoc.

CLEAR-test: combining inference for differential expression and variability in microarray data analysis.

A common goal of microarray experiments is to detect genes that are differentially expressed under distinct experimental conditions. Several statistical tests have been proposed to determine whether the observed changes in gene expression are significant. The t-test assigns a score to each gene on the basis of changes in its expression relative to its estimated variability, in such a way that genes with a higher score (in absolute values) are more likely to be significant. Most variants of the t-test use the complete set of genes to influence the variance estimate for each single gene. However, no inference is made in terms of the variability itself. Here, we highlight the problem of low observed variances in the t-test, when genes with relatively small changes are declared differentially expressed. Alternatively, the z-test could be used although, unlike the t-test, it can declare differentially expressed genes with high observed variances. To overcome this, we propose to combine the z-test, which focuses on large changes, with a chi(2) test to evaluate variability. We call this procedure CLEAR-test and we provide a combined p-value that offers a compromise between both aspects. Analysis of three publicly available microarray datasets reveals the greater performance of the CLEAR-test relative to the t-test and alternative methods. Finally, empirical and simulated data analyses demonstrate the greater reproducibility and statistical power of the CLEAR-test and z-test with respect to current alternative methods. In addition, the CLEAR-test improves the z-test by capturing reproducible genes with high variability.