Epigenetic and transcriptomic consequences of excess X-chromosome material in 47,XXX syndrome-A comparison with Turner syndrome and 46,XX females.

47,XXX (triple X) and Turner syndrome (45,X) are sex chromosomal abnormalities with detrimental effects on health with increased mortality and morbidity. In karyotypical normal females, X-chromosome inactivation balances gene expression between sexes and upregulation of the X chromosome in both sexes maintain stoichiometry with the autosomes. In 47,XXX and Turner syndrome a gene dosage imbalance may ensue from increased or decreased expression from the genes that escape X inactivation, as well as from incomplete X chromosome inactivation in 47,XXX. We aim to study genome-wide DNA-methylation and RNA-expression changes can explain phenotypic traits in 47,XXX syndrome. We compare DNA-methylation and RNA-expression data derived from white blood cells of seven women with 47,XXX syndrome, with data from seven female controls, as well as with seven women with Turner syndrome (45,X). To address these questions, we explored genome-wide DNA-methylation and transcriptome data in blood from seven females with 47,XXX syndrome, seven females with Turner syndrome, and seven karyotypically normal females (46,XX). Based on promoter methylation, we describe a demethylation of six X-chromosomal genes (AMOT, HTR2C, IL1RAPL2, STAG2, TCEANC, ZNF673), increased methylation for GEMIN8, and four differentially methylated autosomal regions related to four genes (SPEG, MUC4, SP6, and ZNF492). We illustrate how these changes seem compensated at the transcriptome level although several genes show differential exon usage. In conclusion, our results suggest an impact of the supernumerary X chromosome in 47,XXX syndrome on the methylation status of selected genes despite an overall comparable expression profile.

Analyses of non-coding somatic drivers in 2,658 cancer whole genomes.

The discovery of drivers of cancer has traditionally focused on protein-coding genes1-4. Here we present analyses of driver point mutations and structural variants in non-coding regions across 2,658 genomes from the Pan-Cancer Analysis of Whole Genomes (PCAWG) Consortium5 of the International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA). For point mutations, we developed a statistically rigorous strategy for combining significance levels from multiple methods of driver discovery that overcomes the limitations of individual methods. For structural variants, we present two methods of driver discovery, and identify regions that are significantly affected by recurrent breakpoints and recurrent somatic juxtapositions. Our analyses confirm previously reported drivers6,7, raise doubts about others and identify novel candidates, including point mutations in the 5' region of TP53, in the 3' untranslated regions of NFKBIZ and TOB1, focal deletions in BRD4 and rearrangements in the loci of AKR1C genes. We show that although point mutations and structural variants that drive cancer are less frequent in non-coding genes and regulatory sequences than in protein-coding genes, additional examples of these drivers will be found as more cancer genomes become available.

DNA hypermethylation and differential gene expression associated with Klinefelter syndrome.

Klinefelter syndrome (KS) has a prevalence ranging from 85 to 250 per 100.000 newborn boys making it the most frequent sex chromosome aneuploidy in the general population. The molecular basis for the phenotypic traits and morbidity in KS are not clarified. We performed genome-wide DNA methylation profiling of leucocytes from peripheral blood samples from 67 KS patients, 67 male controls and 33 female controls, in addition to genome-wide RNA-sequencing profiling in a subset of 9 KS patients, 9 control males and 13 female controls. Characterization of the methylome as well as the transcriptome of both coding and non-coding genes identified a unique epigenetic and genetic landscape of both autosomal chromosomes as well as the X chromosome in KS. A subset of genes show significant correlation between methylation values and expression values. Gene set enrichment analysis of differentially methylated positions yielded terms associated with well-known comorbidities seen in KS. In addition, differentially expressed genes revealed enrichment for genes involved in the immune system, wnt-signaling pathway and neuron development. Based on our data we point towards new candidate genes, which may be implicated in the phenotype and further point towards non-coding genes, which may be involved in X chromosome inactivation in KS.

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

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

Pan-cancer screen for mutations in non-coding elements with conservation and cancer specificity reveals correlations with expression and survival.

Cancer develops by accumulation of somatic driver mutations, which impact cellular function. Mutations in non-coding regulatory regions can now be studied genome-wide and further characterized by correlation with gene expression and clinical outcome to identify driver candidates. Using a new two-stage procedure, called ncDriver, we first screened 507 ICGC whole-genomes from 10 cancer types for non-coding elements, in which mutations are both recurrent and have elevated conservation or cancer specificity. This identified 160 significant non-coding elements, including the TERT promoter, a well-known non-coding driver element, as well as elements associated with known cancer genes and regulatory genes (e.g., PAX5, TOX3, PCF11, MAPRE3). However, in some significant elements, mutations appear to stem from localized mutational processes rather than recurrent positive selection in some cases. To further characterize the driver potential of the identified elements and shortlist candidates, we identified elements where presence of mutations correlated significantly with expression levels (e.g., TERT and CDH10) and survival (e.g., CDH9 and CDH10) in an independent set of 505 TCGA whole-genome samples. In a larger pan-cancer set of 4128 TCGA exomes with expression profiling, we identified mutational correlation with expression for additional elements (e.g., near GATA3, CDC6, ZNF217, and CTCF transcription factor binding sites). Survival analysis further pointed to MIR122, a known marker of poor prognosis in liver cancer. In conclusion, the screen for significant mutation patterns coupled with correlative mutational analysis identified new individual driver candidates and suggest that some non-coding mutations recurrently affect expression and play a role in cancer development.

Non-coding cancer driver candidates identified with a sample- and position-specific model of the somatic mutation rate.

Non-coding mutations may drive cancer development. Statistical detection of non-coding driver regions is challenged by a varying mutation rate and uncertainty of functional impact. Here, we develop a statistically founded non-coding driver-detection method, ncdDetect, which includes sample-specific mutational signatures, long-range mutation rate variation, and position-specific impact measures. Using ncdDetect, we screened non-coding regulatory regions of protein-coding genes across a pan-cancer set of whole-genomes (n = 505), which top-ranked known drivers and identified new candidates. For individual candidates, presence of non-coding mutations associates with altered expression or decreased patient survival across an independent pan-cancer sample set (n = 5454). This includes an antigen-presenting gene (CD1A), where 5'UTR mutations correlate significantly with decreased survival in melanoma. Additionally, mutations in a base-excision-repair gene (SMUG1) correlate with a C-to-T mutational-signature. Overall, we find that a rich model of mutational heterogeneity facilitates non-coding driver identification and integrative analysis points to candidates of potential clinical relevance.

Mutational context and diverse clonal development in early and late bladder cancer.

Bladder cancer (or urothelial cell carcinoma [UCC]) is characterized by field disease (malignant alterations in surrounding mucosa) and frequent recurrences. Whole-genome, exome, and transcriptome sequencing of 38 tumors, including four metachronous tumor pairs and 20 superficial tumors, identified an APOBEC mutational signature in one-third. This was biased toward the sense strand, correlated with mean expression level, and clustered near breakpoints. A>G mutations were up to eight times more frequent on the sense strand (p<0.002) in [ACG]AT contexts. The patient-specific APOBEC signature was negatively correlated to repair-gene expression and was not related to clinicopathological parameters. Mutations in gene families and single genes were related to tumor stage, and expression of chromatin modifiers correlated with survival. Evolutionary and subclonal analyses of early/late tumor pairs showed a unitary origin, and discrete tumor clones contained mutated cancer genes. The ancestral clones contained Pik3ca/Kdm6a mutations and may reflect the field-disease mutations shared among later tumors.

Pairwise comparisons of ten porcine tissues identify differential transcriptional regulation at the gene, isoform, promoter and transcription start site level.

The transcriptome is the absolute set of transcripts in a tissue or cell at the time of sampling. In this study RNA-Seq is employed to enable the differential analysis of the transcriptome profile for ten porcine tissues in order to evaluate differences between the tissues at the gene and isoform expression level, together with an analysis of variation in transcription start sites, promoter usage, and splicing. Totally, 223 million RNA fragments were sequenced leading to the identification of 59,930 transcribed gene locations and 290,936 transcript variants using Cufflinks with similarity to approximately 13,899 annotated human genes. Pairwise analysis of tissues for differential expression at the gene level showed that the smallest differences were between tissues originating from the porcine brain. Interestingly, the relative level of differential expression at the isoform level did generally not vary between tissue contrasts. Furthermore, analysis of differential promoter usage between tissues, revealed a proportionally higher variation between cerebellum (CBE) versus frontal cortex and cerebellum versus hypothalamus (HYP) than in the remaining comparisons. In addition, the comparison of differential transcription start sites showed that the number of these sites is generally increased in comparisons including hypothalamus in contrast to other pairwise assessments. A comprehensive analysis of one of the tissue contrasts, i.e. cerebellum versus heart for differential variation at the gene, isoform, and transcription start site (TSS), and promoter level showed that several of the genes differed at all four levels. Interestingly, these genes were mainly annotated to the "electron transport chain" and neuronal differentiation, emphasizing that "tissue important" genes are regulated at several levels. Furthermore, our analysis shows that the "across tissue approach" has a promising potential when screening for possible explanations for variations, such as those observed at the gene expression levels.

A genome-wide association scan in pig identifies novel regions associated with feed efficiency trait.

Feed conversion ratio (FCR) is an economically important trait in pigs, and feed accounts for a significant proportion of the costs involved in pig production. In this study we used a high-density SNP chip panel, Porcine SNP60 BeadChip, to identify the association between FCR and SNP markers and to study the genetic architecture of the trait. After quality control, a total of 30,847 SNP that could be mapped to the 18 porcine autosomes (SSC) using the pig genome assembly 10.2 were used in the analyses. Deregressed estimated breeding value was used as the response variable. A total of 3,071 Duroc pigs had both FCR data and genotype data. The linkage disequilibrium (r(2)) between adjacent markers was 0.56. Two association mapping approaches were used: a linear mixed model (LMM) based on single-locus regression analysis and a Bayesian variable selection approach (BVS). A total of 79 significant (P < 0.0001) SNP associations on 6 chromosomes were identified by LMM analyses. Out of these, 10 SNP crossed the genome-wide significance threshold. These 10 SNP were all located on SSC 4 and 14. In the BVS analysis, a total of 44 SNP located on 12 chromosomes had posterior probability more than or equal to 0.05 (i.e., Bayes factor ≥ 10). Thirteen SNP were identified by both LMM and BVS. These 13 SNP were located on 4 chromosomes: SSC 4, 7, 8, and 14. Hypoxia inducible factor 1, alpha subunit inhibitor (HIF1AN) and ladybird homeobox 1 (LBX1) are 2 possible candidate genes affecting FCR on SSC 4 and 14, respectively. The study provides a list of SNP associated with FCR and also offers valuable information on the genetic architecture and candidate genes for this trait.

Gene expression profiling of porcine skeletal muscle in the early recovery phase following acute physical activity.

Acute physical activity elicits changes in gene expression in skeletal muscles to promote metabolic changes and to repair exercise-induced muscle injuries. In the present time-course study, pigs were submitted to an acute bout of treadmill running until near exhaustion to determine the impact of unaccustomed exercise on global transcriptional profiles in porcine skeletal muscles. Using a combined microarray and candidate gene approach, we identified a suite of genes that are differentially expressed in muscles during postexercise recovery. Several members of the heat shock protein family and proteins associated with proteolytic events, such as the muscle-specific E3 ubiquitin ligase atrogin-1, were significantly upregulated, suggesting that protein breakdown, prevention of protein aggregation and stabilization of unfolded proteins are important processes for restoration of cellular homeostasis. We also detected an upregulation of genes that are associated with muscle cell proliferation and differentiation, including MUSTN1, ASB5 and CSRP3, possibly reflecting activation, differentiation and fusion of satellite cells to facilitate repair of muscle damage. In addition, exercise increased expression of the orphan nuclear hormone receptor NR4A3, which regulates metabolic functions associated with lipid, carbohydrate and energy homeostasis. Finally, we observed an unanticipated induction of the long non-coding RNA transcript NEAT1, which has been implicated in RNA processing and nuclear retention of adenosine-to-inosine edited mRNAs in the ribonucleoprotein bodies called paraspeckles. These findings expand the complexity of pathways affected by acute contractile activity of skeletal muscle, contributing to a better understanding of the molecular processes that occur in muscle tissue in the recovery phase.

Regulation of basal resistance by a powdery mildew-induced cysteine-rich receptor-like protein kinase in barley.

The receptor-like protein kinases (RLKs) constitute a large and diverse group of proteins controlling numerous plant physiological processes, including development, hormone perception and stress responses. The cysteine-rich RLKs (CRKs) represent a prominent subfamily of transmembrane-anchored RLKs. We have identified a putative barley (Hordeum vulgare) CRK gene family member, designated HvCRK1. The mature putative protein comprises 645 amino acids, and includes a putative receptor domain containing two characteristic 'domain 26 of unknown function' (duf26) domains in the N-terminal region, followed by a rather short 17-amino-acid transmembrane domain, which includes an AAA motif, two features characteristic of endoplasmic reticulum (ER)-targeted proteins and, finally, a characteristic putative protein kinase domain in the C-terminus. The HvCRK1 transcript was isolated from leaves inoculated with the biotrophic fungal pathogen Blumeria graminis f.sp. hordei (Bgh). HvCRK1 transcripts were observed to accumulate transiently following Bgh inoculation of susceptible barley. Transient silencing of HvCRK1 expression in bombarded epidermal cells led to enhanced resistance to Bgh, but did not affect R-gene-mediated resistance. Silencing of HvCRK1 phenocopied the effective penetration resistance found in mlo-resistant barley plants, and the possible link between HvCRK1 and MLO was substantiated by the fact that HvCRK1 induction on Bgh inoculation was dependent on Mlo. Finally, using both experimental and in silico approaches, we demonstrated that HvCRK1 localizes to the ER of barley cells. The negative effect on basal resistance against Bgh and the functional aspects of MLO- and ER-localized HvCRK1 signalling on Bgh inoculation are discussed.

Using microarrays to identify positional candidate genes for QTL: the case study of ACTH response in pigs.

BACKGROUND: Microarray studies can supplement QTL studies by suggesting potential candidate genes in the QTL regions, which by themselves are too large to provide a limited selection of candidate genes. Here we provide a case study where we explore ways to integrate QTL data and microarray data for the pig, which has only a partial genome sequence. We outline various procedures to localize differentially expressed genes on the pig genome and link this with information on published QTL. The starting point is a set of 237 differentially expressed cDNA clones in adrenal tissue from two pig breeds, before and after treatment with adrenocorticotropic hormone (ACTH). RESULTS: Different approaches to localize the differentially expressed (DE) genes to the pig genome showed different levels of success and a clear lack of concordance for some genes between the various approaches. For a focused analysis on 12 genes, overlapping QTL from the public domain were presented. Also, differentially expressed genes underlying QTL for ACTH response were described. Using the latest version of the draft sequence, the differentially expressed genes were mapped to the pig genome. This enabled co-location of DE genes and previously studied QTL regions, but the draft genome sequence is still incomplete and will contain many errors. A further step to explore links between DE genes and QTL at the pathway level was largely unsuccessful due to the lack of annotation of the pig genome. This could be improved by further comparative mapping analyses but this would be time consuming. CONCLUSION: This paper provides a case study for the integration of QTL data and microarray data for a species with limited genome sequence information and annotation. The results illustrate the challenges that must be addressed but also provide a roadmap for future work that is applicable to other non-model species.

Methods for interpreting lists of affected genes obtained in a DNA microarray experiment.

BACKGROUND: The aim of this paper was to describe and compare the methods used and the results obtained by the participants in a joint EADGENE (European Animal Disease Genomic Network of Excellence) and SABRE (Cutting Edge Genomics for Sustainable Animal Breeding) workshop focusing on post analysis of microarray data. The participating groups were provided with identical lists of microarray probes, including test statistics for three different contrasts, and the normalised log-ratios for each array, to be used as the starting point for interpreting the affected probes. The data originated from a microarray experiment conducted to study the host reactions in broilers occurring shortly after a secondary challenge with either a homologous or heterologous species of Eimeria. RESULTS: Several conceptually different analytical approaches, using both commercial and public available software, were applied by the participating groups. The following tools were used: Ingenuity Pathway Analysis, MAPPFinder, LIMMA, GOstats, GOEAST, GOTM, Globaltest, TopGO, ArrayUnlock, Pathway Studio, GIST and AnnotationDbi. The main focus of the approaches was to utilise the relation between probes/genes and their gene ontology and pathways to interpret the affected probes/genes. The lack of a well-annotated chicken genome did though limit the possibilities to fully explore the tools. The main results from these analyses showed that the biological interpretation is highly dependent on the statistical method used but that some common biological conclusions could be reached. CONCLUSION: It is highly recommended to test different analytical methods on the same data set and compare the results to obtain a reliable biological interpretation of the affected genes in a DNA microarray experiment.

A robust linkage map of the porcine autosomes based on gene-associated SNPs.

BACKGROUND: Genetic linkage maps are necessary for mapping of mendelian traits and quantitative trait loci (QTLs). To identify the actual genes, which control these traits, a map based on gene-associated single nucleotide polymorphism (SNP) markers is highly valuable. In this study, the SNPs were genotyped in a large family material comprising more than 5,000 piglets derived from 12 Duroc boars crossed with 236 Danish Landrace/Danish Large White sows. The SNPs were identified in sequence alignments of 4,600 different amplicons obtained from the 12 boars and containing coding regions of genes derived from expressed sequence tags (ESTs) and genomic shotgun sequences. RESULTS: Linkage maps of all 18 porcine autosomes were constructed based on 456 gene-associated and six porcine EST-based SNPs. The total length of the averaged-sex whole porcine autosome was estimated to 1,711.8 cM resulting in an average SNP spacing of 3.94 cM. The female and male maps were estimated to 2,336.1 and 1,441.5 cM, respectively. The gene order was validated through comparisons to the cytogenetic and/or physical location of 203 genes, linkage to evenly spaced microsatellite markers as well as previously reported conserved synteny. A total of 330 previously unmapped genes and ESTs were mapped to the porcine autosome while ten genes were mapped to unexpected locations. CONCLUSION: The linkage map presented here shows high accuracy in gene order. The pedigree family network as well as the large amount of meiotic events provide good reliability and make this map suitable for QTL and association studies. In addition, the linkage to the RH-map of microsatellites makes it suitable for comparison to other QTL studies.

Transcriptomic and proteomic profiling of two porcine tissues using high-throughput technologies.

BACKGROUND: The recent development within high-throughput technologies for expression profiling has allowed for parallel analysis of transcriptomes and proteomes in biological systems such as comparative analysis of transcript and protein levels of tissue regulated genes. Until now, such studies of have only included microarray or short length sequence tags for transcript profiling. Furthermore, most comparisons of transcript and protein levels have been based on absolute expression values from within the same tissue and not relative expression values based on tissue ratios. RESULTS: Presented here is a novel study of two porcine tissues based on integrative analysis of data from expression profiling of identical samples using cDNA microarray, 454-sequencing and iTRAQ-based proteomics. Sequence homology identified 2.541 unique transcripts that are detectable by both microarray hybridizations and 454-sequencing of 1.2 million cDNA tags. Both transcript-based technologies showed high reproducibility between sample replicates of the same tissue, but the correlation across these two technologies was modest. Thousands of genes being differentially expressed were identified with microarray. Out of the 306 differentially expressed genes, identified by 454-sequencing, 198 (65%) were also found by microarray. The relationship between the regulation of transcript and protein levels was analyzed by integrating iTRAQ-based proteomics data. Protein expression ratios were determined for 354 genes, of which 148 could be mapped to both microarray and 454-sequencing data. A comparison of the expression ratios from the three technologies revealed that differences in transcript and protein levels across heart and muscle tissues are positively correlated. CONCLUSION: We show that the reproducibility within cDNA microarray and 454-sequencing is high, but that the agreement across these two technologies is modest. We demonstrate that the regulation of transcript and protein levels across identical tissue samples is positively correlated when the tissue expression ratios are used for comparison. The results presented are of interest in systems biology research in terms of integration and analysis of high-throughput expression data from mammalian tissues.

Gene expression profiles in liver of pigs with extreme high and low levels of androstenone.

BACKGROUND: Boar taint is the unpleasant odour and flavour of the meat of uncastrated male pigs that is primarily caused by high levels of androstenone and skatole in adipose tissue. Androstenone is a steroid and its levels are mainly genetically determined. Studies on androstenone metabolism have, however, focused on a limited number of genes. Identification of additional genes influencing levels of androstenone may facilitate implementation of marker assisted breeding practices. In this study, microarrays were used to identify differentially expressed genes and pathways related to androstenone metabolism in the liver from boars with extreme levels of androstenone in adipose tissue. RESULTS: Liver tissue samples from 58 boars of the two breeds Duroc and Norwegian Landrace, 29 with extreme high and 29 with extreme low levels of androstenone, were selected from more than 2500 individuals. The samples were hybridised to porcine cDNA microarrays and the 1% most significant differentially expressed genes were considered significant. Among the differentially expressed genes were metabolic phase I related genes belonging to the cytochrome P450 family and the flavin-containing monooxygenase FMO1. Additionally, phase II conjugation genes including UDP-glucuronosyltransferases UGT1A5, UGT2A1 and UGT2B15, sulfotransferase STE, N-acetyltransferase NAT12 and glutathione S-transferase were identified. Phase I and phase II metabolic reactions increase the water solubility of steroids and play a key role in their elimination. Differential expression was also found for genes encoding 17beta-hydroxysteroid dehydrogenases (HSD17B2, HSD17B4, HSD17B11 and HSD17B13) and plasma proteins alpha-1-acid glycoprotein (AGP) and orosomucoid (ORM1). 17beta-hydroxysteroid dehydrogenases and plasma proteins regulate the availability of steroids by controlling the amount of active steroids accessible to receptors and available for metabolism. Differences in the expression of FMO1, NAT12, HSD17B2 and HSD17B13 were verified by quantitative real competitive PCR. CONCLUSION: A number of genes and pathways related to metabolism of androstenone in liver were identified, including new candidate genes involved in phase I oxidation metabolism, phase II conjugation metabolism, and regulation of steroid availability. The study is a first step towards a deeper understanding of enzymes and regulators involved in pathways of androstenone metabolism and may ultimately lead to the discovery of markers to reduce boar taint.

The EADGENE Microarray Data Analysis Workshop (open access publication).

Microarray analyses have become an important tool in animal genomics. While their use is becoming widespread, there is still a lot of ongoing research regarding the analysis of microarray data. In the context of a European Network of Excellence, 31 researchers representing 14 research groups from 10 countries performed and discussed the statistical analyses of real and simulated 2-colour microarray data that were distributed among participants. The real data consisted of 48 microarrays from a disease challenge experiment in dairy cattle, while the simulated data consisted of 10 microarrays from a direct comparison of two treatments (dye-balanced). While there was broader agreement with regards to methods of microarray normalisation and significance testing, there were major differences with regards to quality control. The quality control approaches varied from none, through using statistical weights, to omitting a large number of spots or omitting entire slides. Surprisingly, these very different approaches gave quite similar results when applied to the simulated data, although not all participating groups analysed both real and simulated data. The workshop was very successful in facilitating interaction between scientists with a diverse background but a common interest in microarray analyses.

Analysis of the real EADGENE data set: multivariate approaches and post analysis (open access publication).

The aim of this paper was to describe, and when possible compare, the multivariate methods used by the participants in the EADGENE WP1.4 workshop. The first approach was for class discovery and class prediction using evidence from the data at hand. Several teams used hierarchical clustering (HC) or principal component analysis (PCA) to identify groups of differentially expressed genes with a similar expression pattern over time points and infective agent (E. coli or S. aureus). The main result from these analyses was that HC and PCA were able to separate tissue samples taken at 24 h following E. coli infection from the other samples. The second approach identified groups of differentially co-expressed genes, by identifying clusters of genes highly correlated when animals were infected with E. coli but not correlated more than expected by chance when the infective pathogen was S. aureus. The third approach looked at differential expression of predefined gene sets. Gene sets were defined based on information retrieved from biological databases such as Gene Ontology. Based on these annotation sources the teams used either the GlobalTest or the Fisher exact test to identify differentially expressed gene sets. The main result from these analyses was that gene sets involved in immune defence responses were differentially expressed.

Analysis of the real EADGENE data set: comparison of methods and guidelines for data normalisation and selection of differentially expressed genes (open access publication).

A large variety of methods has been proposed in the literature for microarray data analysis. The aim of this paper was to present techniques used by the EADGENE (European Animal Disease Genomics Network of Excellence) WP1.4 participants for data quality control, normalisation and statistical methods for the detection of differentially expressed genes in order to provide some more general data analysis guidelines. All the workshop participants were given a real data set obtained in an EADGENE funded microarray study looking at the gene expression changes following artificial infection with two different mastitis causing bacteria: Escherichia coli and Staphylococcus aureus. It was reassuring to see that most of the teams found the same main biological results. In fact, most of the differentially expressed genes were found for infection by E. coli between uninfected and 24 h challenged udder quarters. Very little transcriptional variation was observed for the bacteria S. aureus. Lists of differentially expressed genes found by the different research teams were, however, quite dependent on the method used, especially concerning the data quality control step. These analyses also emphasised a biological problem of cross-talk between infected and uninfected quarters which will have to be dealt with for further microarray studies.