Jelle J. Goeman

A global test for groups of genes: testing association with a clinical outcome

MOTIVATION This paper presents a global test to be used for the analysis of microarray data. Using this test it can be determined whether the global expression pattern of a group of genes is significantly related to some clinical outcome of interest. Groups of genes may be any size from a single gene to all genes on the chip (e.g. known pathways, specific areas of the genome or clusters from a cluster analysis). RESULT The test allows groups of genes of different size to be compared, because the test gives one p-value for the group, not a p-value for each gene. Researchers can use the test to investigate hypotheses based on theory or past research or to mine gene ontology databases for interesting pathways. Multiple testing problems do not occur unless many groups are tested. Special attention is given to visualizations of the test result, focussing on the associations between samples and showing the impact of individual genes on the test result. AVAILABILITY An R-package globaltest is available from http://www.bioconductor.org

Analyzing gene expression data in terms of gene sets

MOTIVATION Many statistical tests have been proposed in recent years for analyzing gene expression data in terms of gene sets, usually from Gene Ontology. These methods are based on widely different methodological assumptions. Some approaches test differential expression of each gene set against differential expression of the rest of the genes, whereas others test each gene set on its own. Also, some methods are based on a model in which the genes are the sampling units, whereas others treat the subjects as the sampling units. This article aims to clarify the assumptions behind different approaches and to indicate a preferential methodology of gene set testing. RESULTS We identify some crucial assumptions which are needed by the majority of methods. P-values derived from methods that use a model which takes the genes as the sampling unit are easily misinterpreted, as they are based on a statistical model that does not resemble the biological experiment actually performed. Furthermore, because these models are based on a crucial and unrealistic independence assumption between genes, the P-values derived from such methods can be wildly anti-conservative, as a simulation experiment shows. We also argue that methods that competitively test each gene set against the rest of the genes create an unnecessary rift between single gene testing and gene set testing.

L1 Penalized Estimation in the Cox Proportional Hazards Model

This article presents a novel algorithm that efficiently computes L(1) penalized (lasso) estimates of parameters in high-dimensional models. The lasso has the property that it simultaneously performs variable selection and shrinkage, which makes it very useful for finding interpretable prediction rules in high-dimensional data. The new algorithm is based on a combination of gradient ascent optimization with the Newton-Raphson algorithm. It is described for a general likelihood function and can be applied in generalized linear models and other models with an L(1) penalty. The algorithm is demonstrated in the Cox proportional hazards model, predicting survival of breast cancer patients using gene expression data, and its performance is compared with competing approaches. An R package, penalized, that implements the method, is available on CRAN.

Genetic susceptibility to respiratory syncytial virus bronchiolitis is predominantly associated with innate immune genes

BACKGROUND Respiratory syncytial virus (RSV) is a common cause of severe lower respiratory tract infection in infants. Only a proportion of children infected with RSV require hospitalization. Because known risk factors for severe disease, such as premature birth, cannot fully explain differences in disease severity, genetic factors have been implicated. METHODS To study the complexity of RSV susceptibility and to identify the genes and biological pathways involved in its development, we performed a genetic association study involving 470 children hospitalized for RSV bronchiolitis, their parents, and 1008 random, population controls. We analyzed 384 single-nucleotide polymorphisms (SNPs) in 220 candidate genes involved in airway mucosal responses, innate immunity, chemotaxis, adaptive immunity, and allergic asthma. RESULTS SNPs in the innate immune genes VDR (rs10735810; P=.0017), JUN (rs11688; P=.0093), IFNA5 (rs10757212; P=.0093), and NOS2 (rs1060826; P=.0031) demonstrated the strongest association with bronchiolitis. Apart from association at the allele level, these 4 SNPs also demonstrated association at the genotype level (P=.0056, P=.0285, P=.0372, and P=.0117 for the SNPs in VDR, JUN, IFNA5, and NOS2, respectively). The role of innate immunity as a process was reinforced by association of the whole group of innate immune SNPs when the global test for groups of genes was applied (P=.046). CONCLUSION SNPs in innate immune genes are important in determining susceptibility to RSV bronchiolitis.

DNA Methylation Signatures Link Prenatal Famine Exposure to Growth and Metabolism

Periconceptional diet may persistently influence DNA methylation levels with phenotypic consequences. However, a comprehensive assessment of the characteristics of prenatal malnutrition-associated differentially methylated regions (P-DMRs) is lacking in humans. Here we report on a genome-scale analysis of differential DNA methylation in whole blood after periconceptional exposure to famine during the Dutch Hunger Winter. We show that P-DMRs preferentially occur at regulatory regions, are characterized by intermediate levels of DNA methylation and map to genes enriched for differential expression during early development. Validation and further exploratory analysis of six P-DMRs highlight the critical role of gestational timing. Interestingly, differential methylation of the P-DMRs extends along pathways related to growth and metabolism. P-DMRs located in INSR and CPT1A have enhancer activity in vitro and differential methylation is associated with birth weight and serum LDL cholesterol. Epigenetic modulation of pathways by prenatal malnutrition may promote an adverse metabolic phenotype in later life.

Testing association of a pathway with survival using gene expression data

MOTIVATION A recent surge of interest in survival as the primary clinical endpoint of microarray studies has called for an extension of the Global Test methodology to survival. RESULTS We present a score test for association of the expression profile of one or more groups of genes with a (possibly censored) survival time. Groups of genes may be pathways, areas of the genome, clusters from a cluster analysis or all genes on a chip. The test allows one to test hypotheses about the influence of these groups of genes on survival directly, without the intermediary of single gene testing. The test is based on the Cox proportional hazards model and is calculated using martingale residuals. It is possible to adjust the test for the presence of covariates. We also present a diagnostic graph to assist in the interpretation of the test result, visualizing the influence of genes. The test is applied to a tumor dataset, revealing pathways from the gene ontology database that are associated with survival of patients. AVAILABILITY The Global Test for survival has been incorporated into the R-package globaltest (version 3.0), available at http://www.bioconductor.org

Identification and systematic annotation of tissue-specific differentially methylated regions using the Illumina 450k array

BackgroundDNA methylation has been recognized as a key mechanism in cell differentiation. Various studies have compared tissues to characterize epigenetically regulated genomic regions, but due to differences in study design and focus there still is no consensus as to the annotation of genomic regions predominantly involved in tissue-specific methylation. We used a new algorithm to identify and annotate tissue-specific differentially methylated regions (tDMRs) from Illumina 450k chip data for four peripheral tissues (blood, saliva, buccal swabs and hair follicles) and six internal tissues (liver, muscle, pancreas, subcutaneous fat, omentum and spleen with matched blood samples).ResultsThe majority of tDMRs, in both relative and absolute terms, occurred in CpG-poor regions. Further analysis revealed that these regions were associated with alternative transcription events (alternative first exons, mutually exclusive exons and cassette exons). Only a minority of tDMRs mapped to gene-body CpG islands (13%) or CpG islands shores (25%) suggesting a less prominent role for these regions than indicated previously. Implementation of ENCODE annotations showed enrichment of tDMRs in DNase hypersensitive sites and transcription factor binding sites. Despite the predominance of tissue differences, inter-individual differences in DNA methylation in internal tissues were correlated with those for blood for a subset of CpG sites in a locus- and tissue-specific manner.ConclusionsWe conclude that tDMRs preferentially occur in CpG-poor regions and are associated with alternative transcription. Furthermore, our data suggest the utility of creating an atlas cataloguing variably methylated regions in internal tissues that correlate to DNA methylation measured in easy accessible peripheral tissues.

Enhancing scatterplots with smoothed densities

MOTIVATION Scatterplots of microarray data generally contain a very large number of dots, making it difficult to get a good impression of their distribution in dense areas. RESULTS We present a fast and simple algorithm for two-dimensional histogram smoothing, to visually enhance scatterplots. AVAILABILITY Functions for Matlab and R are available from the corresponding author.

Multiple hypothesis testing in genomics

This paper presents an overview of the current state of the art in multiple testing in genomics data from a users perspective. We describe methods for familywise error control, false discovery rate control and false discovery proportion estimation and confidence, both conceptually and practically, and explain when to use which type of error rate. We elaborate on the assumptions underlying the methods and discuss pitfalls in the interpretation of results. In our discussion, we take into account the exploratory nature of genomics experiments, looking at selection of genes before or after testing, and at the role of validation experiments.

Poly(A) binding protein nuclear 1 levels affect alternative polyadenylation

The choice for a polyadenylation site determines the length of the 3′-untranslated region (3′-UTRs) of an mRNA. Inclusion or exclusion of regulatory sequences in the 3′-UTR may ultimately affect gene expression levels. Poly(A) binding protein nuclear 1 (PABPN1) is involved in polyadenylation of pre-mRNAs. An alanine repeat expansion in PABPN1 (exp-PABPN1) causes oculopharyngeal muscular dystrophy (OPMD). We hypothesized that previously observed disturbed gene expression patterns in OPMD muscles may have been the result of an effect of PABPN1 on alternative polyadenylation, influencing mRNA stability, localization and translation. A single molecule polyadenylation site sequencing method was developed to explore polyadenylation site usage on a genome-wide level in mice overexpressing exp-PABPN1. We identified 2012 transcripts with altered polyadenylation site usage. In the far majority, more proximal alternative polyadenylation sites were used, resulting in shorter 3′-UTRs. 3′-UTR shortening was generally associated with increased expression. Similar changes in polyadenylation site usage were observed after knockdown or overexpression of expanded but not wild-type PABPN1 in cultured myogenic cells. Our data indicate that PABPN1 is important for polyadenylation site selection and that reduced availability of functional PABPN1 in OPMD muscles results in use of alternative polyadenylation sites, leading to large-scale deregulation of gene expression.