Comparison of false discovery rate methods in identifying genes with differential expression

Current high-throughput techniques such as microarray in genomics or mass spectrometry in proteomics usually generate thousands of hypotheses to be tested simultaneously. The usual purpose of these techniques is to identify a subset of interesting cases that deserve further investigation. As a consequence, the control of false positives among the tests called "significant" becomes a critical issue for researchers. Over the past few years, several false discovery rate (FDR)-controlling methods have been proposed; each method favors certain scenarios and is introduced with the purpose of improving the control of FDR at the targeted level. In this paper, we compare the performance of the five FDR-controlling methods proposed by Benjamini et al., the qvalue method proposed by Storey, and the traditional Bonferroni method. The purpose is to investigate the "observed" sensitivity of each method on typical microarray experiments in which the majority (or all) of the truth is unknown. Based on two well-studied microarray datasets, it is found that in terms of the "apparent" test power, the ranking of the FDR methods is given as Step-down相似文献   

Interaction-based discovery of functionally important genes in cancers

A major challenge in cancer genomics is uncovering genes with an active role in tumorigenesis from a potentially large pool of mutated genes across patient samples. Here we focus on the interactions that proteins make with nucleic acids, small molecules, ions and peptides, and show that residues within proteins that are involved in these interactions are more frequently affected by mutations observed in large-scale cancer genomic data than are other residues. We leverage this observation to predict genes that play a functionally important role in cancers by introducing a computational pipeline (http://canbind.princeton.edu) for mapping large-scale cancer exome data across patients onto protein structures, and automatically extracting proteins with an enriched number of mutations affecting their nucleic acid, small molecule, ion or peptide binding sites. Using this computational approach, we show that many previously known genes implicated in cancers are enriched in mutations within the binding sites of their encoded proteins. By focusing on functionally relevant portions of proteins—specifically those known to be involved in molecular interactions—our approach is particularly well suited to detect infrequent mutations that may nonetheless be important in cancer, and should aid in expanding our functional understanding of the genomic landscape of cancer.     

Computational method for discovery of estrogen responsive genes

Estrogen has a profound impact on human physiology and affects numerous genes. The classical estrogen reaction is mediated by its receptors (ERs), which bind to the estrogen response elements (EREs) in target gene's promoter region. Due to tedious and expensive experiments, a limited number of human genes are functionally well characterized. It is still unclear how many and which human genes respond to estrogen treatment. We propose a simple, economic, yet effective computational method to predict a subclass of estrogen responsive genes. Our method relies on the similarity of ERE frames across different promoters in the human genome. Matching ERE frames of a test set of 60 known estrogen responsive genes to the collection of over 18,000 human promoters, we obtained 604 candidate genes. Evaluating our result by comparison with the published microarray data and literature, we found that more than half (53.6%, 324/604) of predicted candidate genes are responsive to estrogen. We believe this method can significantly reduce the number of testing potential estrogen target genes and provide functional clues for annotating part of genes that lack functional information.     

Genome-wide discovery of missing genes in biological pathways of prokaryotes

Background

Reconstruction of biological pathways is typically done through mapping well-characterized pathways of model organisms to a target genome, through orthologous gene mapping. A limitation of such pathway-mapping approaches is that the mapped pathway models are constrained by the composition of the template pathways, e.g., some genes in a target pathway may not have corresponding genes in the template pathways, the so-called “missing gene” problem.

Methods

We present a novel pathway-expansion method for identifying additional genes that are possibly involved in a target pathway after pathway mapping, to fill holes caused by missing genes as well as to expand the mapped pathway model. The basic idea of the algorithm is to identify genes in the target genome whose homologous genes share common operons with homologs of any mapped pathway genes in some reference genome, and to add such genes to the target pathway if their functions are consistent with the cellular function of the target pathway.

Results

We have implemented this idea using a graph-theoretic approach and demonstrated the effectiveness of the algorithm on known pathways of E. coli in the KEGG database. On all KEGG pathways containing at least 5 genes, our method achieves an average of 60% positive predictive value (PPV) and the performance is increased with more seed genes added. Analysis shows that our method is highly robust.

Conclusions

An effective method is presented to find missing genes in biological pathways of prokaryotes, which achieves high prediction reliability on E. coli at a genome level. Numerous missing genes are found to be related to knwon E. coli pathways, which can be further validated through biological experiments. Overall this method is robust and can be used for functional inference.

     

Tobacco-smoking-related differential DNA methylation: 27K discovery and replication

Tobacco smoking is responsible for substantial morbidity and mortality worldwide, in particular through cardiovascular, pulmonary, and malignant pathology. CpG methylation might plausibly play a role in a variety of smoking-related phenomena, as suggested by candidate gene promoter or global methylation studies. Arrays allowing hypothesis-free searches on a scale resembling genome-wide studies of SNPs have become available only very recently. Methylation extents in peripheral-blood DNA were assessed at 27,578 sites in more than 14,000 gene promoter regions in 177 current smokers, former smokers, and those who had never smoked, with the use of the Illumina HumanMethylation 27K BeadChip. This revealed a single locus, cg03636183, located in F2RL3, with genome-wide significance for lower methylation in smokers (p = 2.68 × 10(-31)). This was similarly significant in 316 independent replication samples analyzed by mass spectrometry and Sequenom EpiTyper (p = 6.33 × 10(-34)). Our results, which were based on a rigorous replication approach, show that the gene coding for a potential drug target of cardiovascular importance features altered methylation patterns in smokers. To date, this gene had not attracted attention in the literature on smoking.     

Robust discovery of periodically expressed genes using the laplace periodogram

Background  

Time-course gene expression analysis has become important in recent developments due to the increasingly available experimental data. The detection of genes that are periodically expressed is an important step which allows us to study the regulatory mechanisms associated with the cell cycle.     

Quantitating tissue specificity of human genes to facilitate biomarker discovery

We describe a method to identify candidate cancer biomarkers by analyzing numeric approximations of tissue specificity of human genes. These approximations were calculated by analyzing predicted tissue expression distributions of genes derived from mapping expressed sequence tags (ESTs) to the human genome sequence using a binary indexing algorithm. Tissue-specificity values facilitated high-throughput analysis of the human genes and enabled the identification of genes highly specific to different tissues. Tissue expression distributions for several genes were compared to estimates obtained from other public gene expression datasets and experimentally validated using quantitative RT-PCR on RNA isolated from several human tissues. Our results demonstrate that most human genes ( approximately 98%) are expressed in many tissues (low specificity), and only a small number of genes possess very specific tissue expression profiles. These genes comprise a rich dataset from which novel therapeutic targets and novel diagnostic serum biomarkers may be selected.     

Identification of yeast meiosis-specific genes by differential display

Meiosis, a specialized cell division process, occurs in all sexually reproducing organisms. During this process a diploid cell undergoes a single round of DNA replication followed by two rounds of nuclear division to produce four haploid gametes. In yeast, the meiotic products are packaged into four spores that are enclosed in a sac known as an ascus. To enhance our understanding of the meiotic developmental pathway and spore formation, we followed differential expression of genes in meiotic versus vegetatively growing cells in the yeast Saccharomyces cerevisiae. Such comparative analyses have identified five different classes of genes that are expressed at different stages of the sporulation program. We identified several meiosis-specific genes including some already known to be induced during meiosis. Here we describe one of these previously uncharacterized genes, SSP1, which plays an essential role in meiosis and spore formation. SSP1 is induced midway through meiosis, and the homozygous mutant-diploid cells fail to sporulate. In ssp1 cells, meiosis is delayed, nuclei fragment after meiosis II, and viability declines rapidly. The ssp1 defect is not related to a microtubule-cytoskeletal-dependent event and is independent of two rounds of meiotic divisions. Our results suggest that Ssp1 is likely to function in a pathway that controls meiotic nuclear divisions and coordinates meiosis and spore formation. Functional analysis of other uncharacterized genes is underway.     

Cloning oncogenic ras-regulated genes by differential display

The coordinated regulation of gene expression is a key cellular function that specifies cell characteristics as well as controls normal physiological processes of the organism. Deregulation of this gene expression leads to a variety of abnormal conditions such as cancer. The ras oncogene is one of the most frequently found mutations in various types of human cancer. The mutated Ras protein constitutively elicits multiple mitogenic signals to the nucleus to alter gene expression of target genes that are involved in a broad range of normal cellular functions. Thus the identification of these genes may provide an important tool toward the understanding of these pathogenic processes. As a first step to reveal these processes at the molecular level and to dissect the key pathway employed by oncogenic Ras protein, we have looked for its target genes in rodent model cell lines using the differential display method. Our initial screening has isolated a number of genes either up- or downregulated by oncogenic ras activation. Although the functional analyses of these genes in terms of ras-mediated cell transformation will be the major challenge, differential display has come to be a very efficient tool that helped us move to the next step. In this short report, we focus primarily on the technical aspects of differential display and experimental designs used in this study.     

Ranking genes with respect to differential expression

Background  

In the pharmaceutical industry and in academia substantial efforts are made to make the best use of the promising microarray technology. The data generated by microarrays are more complex than most other biological data attracting much attention at this point. A method for finding an optimal test statistic with which to rank genes with respect to differential expression is outlined and tested. At the heart of the method lies an estimate of the false negative and false positive rates. Both investing in false positives and missing true positives lead to a waste of resources. The procedure sets out to minimise these errors. For calculation of the false positive and negative rates a simulation procedure is invoked.     

Divergence and differential expression of soybean actin genes

DNA sequence analysis as well as genomic blotting experiments using cloned soybean actin DNA sequences as probes show that large sequence heterogeneity exists among members of the soybean actin multigene family. This heterogeneity suggested that the members of this family might be diverged in function and/or regulation. Five of the six soybean actin gene family members examined are shown to be significantly more diverged from one another than members of other known actin gene families. This high level of divergence was utilized in the preparation of actin gene-specific probes in the analysis of the complexity and expression of these members of the soybean actin gene family. Hybridization studies indicate that the six soybean actin genes fall into three classes with a pair of genes in each class. These six genes account for all but two actin gene fragments detected in the soybean genome. We have compared the relative steady state mRNA levels of these classes of soybean actin genes in three organs of soybean. We find that actin genes SAc6 and SAc7 are most highly expressed accounting for 80% of all actin mRNA with respect to the six soybean actin genes examined. Actin genes SAc3 and SAc1 are expressed at intermediate and low levels respectively; and SAc2 and SAc4 are expressed at barely detectable levels. Four of the six soybean actin genes appear to be expressed at the same level in root, shoot and hypocotyl. SAc3 and SAc7 genes appear to be more highly expressed in shoot and 2,4-dichlorophenoxyacetic acid-induced hypocotyl than in root and hypocotyl.(ABSTRACT TRUNCATED AT 250 WORDS)     

Biomarker discovery from pancreatic cancer secretome using a differential proteomic approach

Quantitative proteomics can be used as a screening tool for identification of differentially expressed proteins as potential biomarkers for cancers. Candidate biomarkers from such studies can subsequently be tested using other techniques for use in early detection of cancers. Here we demonstrate the use of stable isotope labeling with amino acids in cell culture (SILAC) method to compare the secreted proteins (secretome) from pancreatic cancer-derived cells with that from non-neoplastic pancreatic ductal cells. We identified 145 differentially secreted proteins (>1.5-fold change), several of which were previously reported as either up-regulated (e.g. cathepsin D, macrophage colony stimulation factor, and fibronectin receptor) or down-regulated (e.g. profilin 1 and IGFBP-7) proteins in pancreatic cancer, confirming the validity of our approach. In addition, we identified several proteins that have not been correlated previously with pancreatic cancer including perlecan (HSPG2), CD9 antigen, fibronectin receptor (integrin beta1), and a novel cytokine designated as predicted osteoblast protein (FAM3C). The differential expression of a subset of these novel proteins was validated by Western blot analysis. In addition, overexpression of several proteins not described previously to be elevated in human pancreatic cancer (CD9, perlecan, SDF4, apoE, and fibronectin receptor) was confirmed by immunohistochemical labeling using pancreatic cancer tissue microarrays suggesting that these could be further pursued as potential biomarkers. Lastly the protein expression data from SILAC were compared with mRNA expression data obtained using gene expression microarrays for the two cell lines (Panc1 and human pancreatic duct epithelial), and a correlation coefficient (r) of 0.28 was obtained, confirming previously reported poor associations between RNA and protein expression studies.