Development of computational tools for the integrated analysis of DNA microarray data with applications in cancer research

The MAP-i Doctoral Program of the Universities of Minho, Aveiro and Porto

Integrating data from heterogeneous DNA microarray platforms

DNA microarrays are one of the most used technologies for gene expression measurement. However, there are several distinct microarray platforms, from different manufacturers, each with its own measurement protocol, resulting in data that can hardly be compared or directly integrated. Data integration from multiple sources aims to improve the assertiveness of statistical tests, reducing the data dimensionality problem. The integration of heterogeneous DNA microarray platforms comprehends a set of tasks that range from the re-annotation of the features used on gene expression, to data normalization and batch effect elimination. In this work, a complete methodology for gene expression data integration and application is proposed, which comprehends a transcript-based re-annotation process and several methods for batch effect attenuation. The integrated data will be used to select the best feature set and learning algorithm for a brain tumor classification case study. The integration will consider data from heterogeneous Agilent and Affymetrix platforms, collected from public gene expression databases, such as The Cancer Genome Atlas and Gene Expression Omnibus.

The Interleukin-1 receptor antagonist is a direct target gene of PPARalpha in liver.

BACKGROUND/AIMS: The Peroxisome Proliferator-Activated Receptor (PPAR) alpha belongs to the superfamily of Nuclear Receptors and plays an important role in numerous cellular processes, including lipid metabolism. It is known that PPARalpha also has an anti-inflammatory effect, which is mainly achieved by down-regulating pro-inflammatory genes. The objective of this study was to further characterize the role of PPARalpha in inflammatory gene regulation in liver. RESULTS: According to Affymetrix micro-array analysis, the expression of various inflammatory genes in liver was decreased by treatment of mice with the synthetic PPARalpha agonist Wy14643 in a PPARalpha-dependent manner. In contrast, expression of Interleukin-1 receptor antagonist (IL-1ra), which was acutely stimulated by LPS treatment, was induced by PPARalpha. Up-regulation of IL-1ra by LPS was lower in PPARalpha -/- mice compared to Wt mice. Transactivation and chromatin immunoprecipitation studies identified IL-1ra as a direct positive target gene of PPARalpha with a functional PPRE present in the promoter. Up-regulation of IL-1ra by PPARalpha was conserved in human HepG2 hepatoma cells and the human monocyte/macrophage THP-1 cell line. CONCLUSIONS: In addition to down-regulating expression of pro-inflammatory genes, PPARalpha suppresses the inflammatory response by direct up-regulation of genes with anti-inflammatory properties.

Genome-wide association study of recurrent major depressive disorder in two European case-control cohorts.

Major depressive disorder (MDD) is a highly prevalent disorder with substantial heritability. Heritability has been shown to be substantial and higher in the variant of MDD characterized by recurrent episodes of depression. Genetic studies have thus far failed to identify clear and consistent evidence of genetic risk factors for MDD. We conducted a genome-wide association study (GWAS) in two independent datasets. The first GWAS was performed on 1022 recurrent MDD patients and 1000 controls genotyped on the Illumina 550 platform. The second was conducted on 492 recurrent MDD patients and 1052 controls selected from a population-based collection, genotyped on the Affymetrix 5.0 platform. Neither GWAS identified any SNP that achieved GWAS significance. We obtained imputed genotypes at the Illumina loci for the individuals genotyped on the Affymetrix platform, and performed a meta-analysis of the two GWASs for this common set of approximately half a million SNPs. The meta-analysis did not yield genome-wide significant results either. The results from our study suggest that SNPs with substantial odds ratio are unlikely to exist for MDD, at least in our datasets and among the relatively common SNPs genotyped or tagged by the half-million-loci arrays. Meta-analysis of larger datasets is warranted to identify SNPs with smaller effects or with rarer allele frequencies that contribute to the risk of MDD.

Uncovering hidden duplicated content in public transcriptomics data.

As part of the development of the database Bgee (a dataBase for Gene Expression Evolution), we annotate and analyse expression data from different types and different sources, notably Affymetrix data from GEO and ArrayExpress, and RNA-Seq data from SRA. During our quality control procedure, we have identified duplicated content in GEO and ArrayExpress, affecting ∼14% of our data: fully or partially duplicated experiments from independent data submissions, Affymetrix chips reused in several experiments, or reused within an experiment. We present here the procedure that we have established to filter such duplicates from Affymetrix data, and our procedure to identify future potential duplicates in RNA-Seq data. Database URL: http://bgee.unil.ch/

MIMAS 3.0 is a Multiomics Information Management and Annotation System.

BACKGROUND: DNA sequence integrity, mRNA concentrations and protein-DNA interactions have been subject to genome-wide analyses based on microarrays with ever increasing efficiency and reliability over the past fifteen years. However, very recently novel technologies for Ultra High-Throughput DNA Sequencing (UHTS) have been harnessed to study these phenomena with unprecedented precision. As a consequence, the extensive bioinformatics environment available for array data management, analysis, interpretation and publication must be extended to include these novel sequencing data types. DESCRIPTION: MIMAS was originally conceived as a simple, convenient and local Microarray Information Management and Annotation System focused on GeneChips for expression profiling studies. MIMAS 3.0 enables users to manage data from high-density oligonucleotide SNP Chips, expression arrays (both 3'UTR and tiling) and promoter arrays, BeadArrays as well as UHTS data using MIAME-compliant standardized vocabulary. Importantly, researchers can export data in MAGE-TAB format and upload them to the EBI's ArrayExpress certified data repository using a one-step procedure. CONCLUSION: We have vastly extended the capability of the system such that it processes the data output of six types of GeneChips (Affymetrix), two different BeadArrays for mRNA and miRNA (Illumina) and the Genome Analyzer (a popular Ultra-High Throughput DNA Sequencer, Illumina), without compromising on its flexibility and user-friendliness. MIMAS, appropriately renamed into Multiomics Information Management and Annotation System, is currently used by scientists working in approximately 50 academic laboratories and genomics platforms in Switzerland and France. MIMAS 3.0 is freely available via http://multiomics.sourceforge.net/.

Comparative transcriptomics of rice reveals an ancient pattern of response to microbial colonization.

Glomalean fungi induce and colonize symbiotic tissue called arbuscular mycorrhiza on the roots of most land plants. Other fungi also colonize plants but cause disease not symbiosis. Whole-transcriptome analysis using a custom-designed Affymetrix Gene-Chip and confirmation with real-time RT-PCR revealed 224 genes affected during arbuscular mycorrhizal symbiosis. We compared these transcription profiles with those from rice roots that were colonized by pathogens (Magnaporthe grisea and Fusarium moniliforme). Over 40% of genes showed differential regulation caused by both the symbiotic and at least one of the pathogenic interactions. A set of genes was similarly expressed in all three associations, revealing a conserved response to fungal colonization. The responses that were shared between pathogen and symbiont infection may play a role in compatibility. Likewise, the responses that are different may cause disease. Some of the genes that respond to mycorrhizal colonization may be involved in the uptake of phosphate. Indeed, phosphate addition mimicked the effect of mycorrhiza on 8% of the tested genes. We found that 34% of the mycorrhiza-associated rice genes were also associated with mycorrhiza in dicots, revealing a conserved pattern of response between the two angiosperm classes.

Genome-wide linkage and association analyses to identify genes influencing adiponectin levels: the GEMS Study.

Adiponectin has a variety of metabolic effects on obesity, insulin sensitivity, and atherosclerosis. To identify genes influencing variation in plasma adiponectin levels, we performed genome-wide linkage and association scans of adiponectin in two cohorts of subjects recruited in the Genetic Epidemiology of Metabolic Syndrome Study. The genome-wide linkage scan was conducted in families of Turkish and southern European (TSE, n = 789) and Northern and Western European (NWE, N = 2,280) origin. A whole genome association (WGA) analysis (500K Affymetrix platform) was carried out in a set of unrelated NWE subjects consisting of approximately 1,000 subjects with dyslipidemia and 1,000 overweight subjects with normal lipids. Peak evidence for linkage occurred at chromosome 8p23 in NWE subjects (lod = 3.10) and at chromosome 3q28 near ADIPOQ, the adiponectin structural gene, in TSE subjects (lod = 1.70). In the WGA analysis, the single-nucleotide polymorphisms (SNPs) most strongly associated with adiponectin were rs3774261 and rs6773957 (P < 10(-7)). These two SNPs were in high linkage disequilibrium (r(2) = 0.98) and located within ADIPOQ. Interestingly, our fourth strongest region of association (P < 2 x 10(-5)) was to an SNP within CDH13, whose protein product is a newly identified receptor for high-molecular-weight species of adiponectin. Through WGA analysis, we confirmed previous studies showing SNPs within ADIPOQ to be strongly associated with variation in adiponectin levels and further observed these to have the strongest effects on adiponectin levels throughout the genome. We additionally identified a second gene (CDH13) possibly influencing variation in adiponectin levels. The impact of these SNPs on health and disease has yet to be determined.

A genome-wide association search for type 2 diabetes genes in African Americans.

African Americans are disproportionately affected by type 2 diabetes (T2DM) yet few studies have examined T2DM using genome-wide association approaches in this ethnicity. The aim of this study was to identify genes associated with T2DM in the African American population. We performed a Genome Wide Association Study (GWAS) using the Affymetrix 6.0 array in 965 African-American cases with T2DM and end-stage renal disease (T2DM-ESRD) and 1029 population-based controls. The most significant SNPs (n = 550 independent loci) were genotyped in a replication cohort and 122 SNPs (n = 98 independent loci) were further tested through genotyping three additional validation cohorts followed by meta-analysis in all five cohorts totaling 3,132 cases and 3,317 controls. Twelve SNPs had evidence of association in the GWAS (P<0.0071), were directionally consistent in the Replication cohort and were associated with T2DM in subjects without nephropathy (P<0.05). Meta-analysis in all cases and controls revealed a single SNP reaching genome-wide significance (P<2.5×10(-8)). SNP rs7560163 (P = 7.0×10(-9), OR (95% CI) = 0.75 (0.67-0.84)) is located intergenically between RND3 and RBM43. Four additional loci (rs7542900, rs4659485, rs2722769 and rs7107217) were associated with T2DM (P<0.05) and reached more nominal levels of significance (P<2.5×10(-5)) in the overall analysis and may represent novel loci that contribute to T2DM. We have identified novel T2DM-susceptibility variants in the African-American population. Notably, T2DM risk was associated with the major allele and implies an interesting genetic architecture in this population. These results suggest that multiple loci underlie T2DM susceptibility in the African-American population and that these loci are distinct from those identified in other ethnic populations.

Integrative analysis of RUNX1 downstream pathways and target genes.

BACKGROUND: The RUNX1 transcription factor gene is frequently mutated in sporadic myeloid and lymphoid leukemia through translocation, point mutation or amplification. It is also responsible for a familial platelet disorder with predisposition to acute myeloid leukemia (FPD-AML). The disruption of the largely unknown biological pathways controlled by RUNX1 is likely to be responsible for the development of leukemia. We have used multiple microarray platforms and bioinformatic techniques to help identify these biological pathways to aid in the understanding of why RUNX1 mutations lead to leukemia. RESULTS: Here we report genes regulated either directly or indirectly by RUNX1 based on the study of gene expression profiles generated from 3 different human and mouse platforms. The platforms used were global gene expression profiling of: 1) cell lines with RUNX1 mutations from FPD-AML patients, 2) over-expression of RUNX1 and CBFbeta, and 3) Runx1 knockout mouse embryos using either cDNA or Affymetrix microarrays. We observe that our datasets (lists of differentially expressed genes) significantly correlate with published microarray data from sporadic AML patients with mutations in either RUNX1 or its cofactor, CBFbeta. A number of biological processes were identified among the differentially expressed genes and functional assays suggest that heterozygous RUNX1 point mutations in patients with FPD-AML impair cell proliferation, microtubule dynamics and possibly genetic stability. In addition, analysis of the regulatory regions of the differentially expressed genes has for the first time systematically identified numerous potential novel RUNX1 target genes. CONCLUSION: This work is the first large-scale study attempting to identify the genetic networks regulated by RUNX1, a master regulator in the development of the hematopoietic system and leukemia. The biological pathways and target genes controlled by RUNX1 will have considerable importance in disease progression in both familial and sporadic leukemia as well as therapeutic implications

Pregnancy-induced chromatin remodeling in the breast of postmenopausal women.

Early pregnancy and multiparity are known to reduce the risk of women to develop breast cancer at menopause. Based on the knowledge that the differentiation of the breast induced by the hormones of pregnancy plays a major role in this protection, this work was performed with the purpose of identifying what differentiation-associated molecular changes persist in the breast until menopause. Core needle biopsies (CNB) obtained from the breast of 42 nulliparous (NP) and 71 parous (P) postmenopausal women were analyzed in morphology, immunocytochemistry and gene expression. Whereas in the NP breast, nuclei of epithelial cells were large and euchromatic, in the P breast they were small and hyperchromatic, showing strong methylation of histone 3 at lysine 9 and 27. Transcriptomic analysis performed using Affymetrix HG_U133 oligonucleotide arrays revealed that in CNB of the P breast, there were 267 upregulated probesets that comprised genes controlling chromatin organization, transcription regulation, splicing machinery, mRNA processing and noncoding elements including XIST. We concluded that the differentiation process induced by pregnancy is centered in chromatin remodeling and in the mRNA processing reactome, both of which emerge as important regulatory pathways. These are indicative of a safeguard step that maintains the fidelity of the transcription process, becoming the ultimate mechanism mediating the protection of the breast conferred by full-term pregnancy.

Role of ATH5 in the specification of retinal ganglion cells and expression and cell compartmentalization of EFEMP1, a protein associated with Malattia Leventinese

ABSTRACT : The development of the retina is a very complex process, occurring through the progressive restriction of cell fates, from pluripotent cell populations to complex tissues and organs. In all vertebrate species analyzed so far, retinal differentiation starts with the generation of retinal ganglion cells (RGC)s. One of the documented key essential events in the specification of RGCs is the expression of ATHS, an atonal homolog encoding a bHLH transcription factor. Despite the putative role of master regulator of RGC differentiation, the mechanism of integrating its functions into a coherent program underlying the production of this subclass of retinal neurons has not yet been elucidated. By using chromatin immunoprecipitation combined with microarray (ChIP-on-chip) we have screened for ATH5 direct targets in the developing chick retina at two consecutive periods: E3.5 (stage HH22) and E6 (stage HH30), covering the stages of progenitor proliferation, neuroepithelium patterning, RGC specification, cell cycle exit and early neuronal differentiation. In parallel, complementary analysis with Affymetrix expression microarrays was conducted. We compared RGCs versus retina to see if the targets correspond to genes preferentially expressed in RGCs. We also precociously overexpressed ATH5 in the retina of individual embryo, and contralateral retina vas used as a control. Our integrated approach allowed us to establish a compendium of ATH5-targets and enabled us to position ATH5 in the transcription network underlying neurogenesis in the retina. Malattia Leventinese (ML) is an autosomal, dominant retinal dystrophy characterized by extracellular, amorphous deposits known as drusen, between the retinal pigment epithelium (RPE) and Bruch's membrane. On the genetic level, it has been associated with a single missense mutation (R345W) in a widely expressed gene with unknown function called EFEMP1. We determined expression patterns of the EFEMP1 gene in normal and ML human retinas. Our data shown that the upregulation of EFEMP1 is not specific to ML eye, except for the region of the ciliary body. We also analyzed the cell compartmentalization of different versions of the protein (both wild type and mutant). Our studies indicate that both abnormal expression of the EFEMP1 gene and mutation and accumulation of EFEMP 1 protein (inside or outside the cells) might contribute to the ML pathology. Résumé : 1er partie : L'ontogenèse de la rétine est un processus complexe au cours duquel des cellules progénitrices sont engagée, par vagues successives, dans des lignées où elles vont d'abord être déterminées puis vont se différencier pour finalement construire un tissu rétinien composé de cinq classes de neurones (les photorécepteurs, les cellules horizontales, bipolaires, amacrines et ganglionnaires) et d'une seule de cellules gliales (les cellules de Muller). Chez tous les vertébrés, la neurogenèse rétinienne est d'abord marquée par la production des cellules ganglionnaires (RGCs). La production de cette classe de neurone est liée à l'expression du gène ATH5 qui est un homologue du gène atonal chez la Drosophile et qui code pour un facteur de transcription de la famille des protéines basic Helix-Loop-Helix (bHLH). Malgré le rôle central que joue ATH5 dans la production des RGCs, le mécanisme qui intègre la fonction de cette protéine dans le programme de détermination neuronale et ceci en relation avec le développement de la rétine n'est pas encore élucidé. Grâce à une technologie qui permet de combiner la sélection de fragments de chromatine liant ATH5 et la recherche de séquences grâce à des puces d'ADN non-codants (ChIP-on-chip), nous avons recherché des cibles potentielles de la protéine ATH5 dans la rétine en développement. Nous avons conduit cette recherche à deux stades de développement de manière à englober la phase de prolifération cellulaire, la détermination des RGCs, la sortie du cycle cellulaire ainsi que les premières étapes de la différentiation de ces neurones. Des expériences complémentaires nous ont permis de définir les patrons d'expression des gènes sélectionnés ainsi que l'activité promotrice des éléments de régulation identifiés lors de notre criblage. Ces approches expérimentales diverses et complémentaires nous ont permis de répertorier des gènes cibles de la protéine ATH5 et d'établir ainsi des liens fonctionnels entre des voies métaboliques dont nous ne soupçonnions pas jusqu'alors qu'elles puissent être associées à la production d'une classe de neurones centraux. 2ème partie : Malattia Leventinese (ML) est une maladie génétique qui engendre une dystrophie de la rétine. Elle se caractérise par l'accumulation de dépôt amorphe entre l'épithélium pigmentaire et la membrane de Bruch et connu sous le nom de drusen. Cette maladie est liée à une simple mutation non-sens (R345W) dans un gène dénommé EFEMP1 qui est exprimé dans de nombreux tissus mais dont la fonction reste mal définie. Une étude détaillée de l'expression de ce gène dans des rétines humaines a révélé une expression à un niveau élevé du gène EFEMP1 dans divers tissus de l'oeil ML mais également dans des yeux contrôles. Alors que l'accumulation d'ARN messager EFEMP1 dans les cellules de l'épithélium pigmentaire n'est pas spécifique à ML, l'expression de ce gène dans le corps cilié n'a été observée que dans l'oeil ML. Nous avons également comparé la sécrétion de la protéine sauvage avec celle porteuse de la mutation. En résumé, notre étude révèle que le niveau élevé d'expression du gène EFEMP1 ainsi que l'accumulation de la protéine dans certains compartiments cellulaires pourraient contribuer au développement de pathologies rétiniennes liées à ML.

Avaluació de l’estructuració genètica humana a la Península Ibèrica i a d’altres regions del sudoest europeu: implicacions poblacionals i epidemiològiques

Els avenços en tècniques de genotipat de polimorfismes genètics a gran escala estan liderant una revolució en el camp de l’epidemiologia genètica i la genètica de poblacions humanes. La informació aportada per aquestes tècniques ha evidenciat l’existència d’estructuracions poblacionals que poden augmentar l’error en els estudis d’associació a escala genòmica (GWAS, genome-wide association studies). Estudis recents han demostrat la presència d’aquestes estructuracions a nivell interregional i intrarregional a Europa. El present projecte ha avaluat el grau d’estructuració genètica en poblacions de la Península Ibèrica i altres regions del sudoest europeu (Itàlia i França) per quantificar l’impacte que aquesta potencial estructuració pot tenir en el disseny d’estudis d’associació GWAS i reconstruir la història demogràfica de les poblacions de la Mediterrània. Per aconseguir aquests objectius, s’han analitzat mostres de DNA de 770 individus de 26 poblacions de la Península Ibèrica, França, Itàlia i d’altres països de la Mediterrània. Aquestes mostres van ser genotipades per 240000 SNPs utilitzant l’array 250K StyI d’Affymetrix en el marc d’aquest projecte o mitjançant altres arrays d’Affymetrix en els projectes internacionals HapMap i POPRES. S’han realitzat anàlisis estadístiques incloent anàlisis de components principals, Fst, identitat per descendència, desequilibri de lligament, barreres genètiques, etc. Aquests resultats han permés construir un marc de referència de la variabilitat en aquesta regió, avaluar el seu impacte en estudis d’associació i proposar mesures per evitar l’increment de qualsevol tipus d’error (tipus I i II) en estudis nacionals i internacionals. A més, també han permés reconstruir la història de les poblacions humanes de la Mediterrània així com analitzar les seves relacions demogràfiques. Donada la duració limitada d’aquesta acció (24 mesos, d’octubre de 2010 a setembre de 2012), els resultats d’aquest projecte es troben actualment en fase de redacció i conduiran a diverses publicacions en revistes internacionals i a la preparació de comunicacions a congressos.

Identification of molecular apocrine breast tumours by microarray analysis.

Previous microarray studies on breast cancer identified multiple tumour classes, of which the most prominent, named luminal and basal, differ in expression of the oestrogen receptor alpha gene (ER). We report here the identification of a group of breast tumours with increased androgen signalling and a 'molecular apocrine' gene expression profile. Tumour samples from 49 patients with large operable or locally advanced breast cancers were tested on Affymetrix U133A gene expression microarrays. Principal components analysis and hierarchical clustering split the tumours into three groups: basal, luminal and a group we call molecular apocrine. All of the molecular apocrine tumours have strong apocrine features on histological examination (P=0.0002). The molecular apocrine group is androgen receptor (AR) positive and contains all of the ER-negative tumours outside the basal group. Kolmogorov-Smirnov testing indicates that oestrogen signalling is most active in the luminal group, and androgen signalling is most active in the molecular apocrine group. ERBB2 amplification is commoner in the molecular apocrine than the other groups. Genes that best split the three groups were identified by Wilcoxon test. Correlation of the average expression profile of these genes in our data with the expression profile of individual tumours in four published breast cancer studies suggest that molecular apocrine tumours represent 8-14% of tumours in these studies. Our data show that it is possible with microarray data to divide mammary tumour cells into three groups based on steroid receptor activity: luminal (ER+ AR+), basal (ER- AR-) and molecular apocrine (ER- AR+).

Chronic wound healing by fetal cell therapy may be explained by differential gene profiling observed in fetal versus old skin cells.

Engineering of fetal tissue has a high potential for the treatment of acute and chronic wounds of the skin in humans as these cells have high expansion capacity under simple culture conditions and one organ donation can produce Master Cell Banks which can fabricate over 900 million biological bandages (9 x 12cm). In a Phase 1 clinical safety study, cases are presented for the treatment of therapy resistant leg ulcers. All eight patients, representing 13 ulcers, tolerated multiple treatments with fetal biological bandages showing no negative secondary effects and repair processes similar to that seen in 3rd degree burns. Differential gene profiling using Affymetrix gene chips (analyzing 12,500 genes) were accomplished on these banked fetal dermal skin cells compared to banked dermal skin cells of an aged donor in order to point to potential indicators of wound healing. Families of genes involved in cell adhesion and extracellular matrix, cell cycle, cellular signaling, development and immune response show significant differences in regulation between banked fetal and those from banked old skin cells: with approximately 47.0% of genes over-expressed in fetal fibroblasts. It is perhaps these differences which contribute to efficient tissue repair seen in the clinic with fetal cell therapy.