BAMarraytrade mark: Java software for Bayesian analysis of variance for microarray data

Mycobacterium tuberculosis-Induced Bronchoalveolar Lavage Gene Expression Signature in Latent Tuberculosis Infection Is Dominated by Pleiotropic Effects of CD4+ T Cell-Dependent IFN-γ Production despite the Presence of Polyfunctional T Cells within the Airways

Tuberculosis (TB) remains a worldwide public health threat. Development of a more effective vaccination strategy to prevent pulmonary TB, the most common and contagious form of the disease, is a research priority for international TB control. A key to reaching this goal is improved understanding of the mechanisms of local immunity to Mycobacterium tuberculosis, the causative organism of TB. In this study, we evaluated global M. tuberculosis-induced gene expression in airway immune cells obtained by bronchoalveolar lavage (BAL) of individuals with latent TB infection (LTBI) and M. tuberculosis-naive controls. In prior studies, we demonstrated that BAL cells from LTBI individuals display substantial enrichment for M. tuberculosis-responsive CD4⁺ T cells compared with matched peripheral blood samples. We therefore specifically assessed the impact of the depletion of CD4⁺ and CD8⁺ T cells on M. tuberculosis-induced BAL cell gene expression in LTBI. Our studies identified 12 canonical pathways and a 47-gene signature that was both sensitive and specific for the contribution of CD4⁺ T cells to local recall responses to M. tuberculosis In contrast, depletion of CD8⁺ cells did not identify any genes that fit our strict criteria for inclusion in this signature. Although BAL CD4⁺ T cells in LTBI displayed polyfunctionality, the observed gene signature predominantly reflected the impact of IFN-γ production on a wide range of host immune responses. These findings provide a standard for comparison of the efficacy of standard bacillus Calmette-Guérin vaccination as well as novel TB vaccines now in development at impacting the initial response to re-exposure to M. tuberculosis in the human lung.

Transcriptome signature for dietary fructose-specific changes in rat renal cortex: A quantitative approach to physiological relevance

Fructose consumption causes metabolic diseases and renal injury primarily in the renal cortex where fructose is metabolized. Analyzing gene differential expression induced by dietary manipulation is challenging. The effects may depend on the base diet and primary changes likely induce secondary or higher order changes that are difficult to capture by conventional univariate transcriptome analyses. We hypothesized that dietary fructose induces a genetic program in the kidney cortex that favors lipogenesis and gluconeogenesis. To test this, we analyzed renal cortical transcriptomes of rats on normal- and high-salt base diets supplemented with fructose. Both sets of data were analyzed using the Characteristic Direction method to yield fructose-induced gene vectors of associated differential expression values. A fructose-specific "signature" of 139 genes differentially expressed was extracted from the 2 diet vectors by a new algorithm that takes into account a gene's rank and standard deviation of its differential expression value. Of these genes, 97 were annotated and the top 34 accounted for 80% of the signal in the annotated signature. The genes were predominantly proximal tubule-specific, coding for metabolic enzymes or transporters. Cosine similarity of signature genes in the two fructose-induced vectors was >0.78. These 139 genes of the fructose signature contributed 27% and 38% of total differential expression on normal- and high- salt diet, respectively. Principal Component Analysis showed that the individual animals could be grouped according to diet. The fructose signature contained a greater enrichment of Gene Ontology processes related to nutrition and metabolism of fructose than two univariate analysis methods. The major feature of the fructose signature is a change in metabolic programs of the renal proximal tubule consistent with gluconeogenesis and de-novo lipogenesis. This new "signature" constitutes a new metric to bridge the gap between physiological phenomena and differential expression profile.

Identification of Novel Signal Transduction, Immune Function, and Oxidative Stress Genes and Pathways by Topiramate for Treatment of Methamphetamine Dependence Based on Secondary Outcomes

BACKGROUND

Topiramate (TPM) is suggested to be a promising medication for treatment of methamphetamine (METH) dependence, but the molecular basis remains to be elucidated.

METHODS

Among 140 METH-dependent participants randomly assigned to receive either TPM (N = 69) or placebo (N = 71) in a previously conducted randomized controlled trial, 50 TPM- and 49 placebo-treated participants had a total 212 RNA samples available at baseline, week 8, and week 12 time points. Following our primary analysis of gene expression data, we reanalyzed the microarray expression data based on a latent class analysis of binary secondary outcomes during weeks 1-12 that provided a classification of 21 responders and 31 non-responders with consistent responses at both time points.

RESULTS

Based on secondary outcomes, 1,381, 576, 905, and 711 differentially expressed genes at nominal P values < 0.05 were identified in responders versus non-responders for week 8 TPM, week 8 placebo, week 12 TPM, and week 12 placebo groups, respectively. Among 1,381 genes identified in week 8 TPM responders, 359 genes were identified in both week 8 and week 12 TPM groups, of which 300 genes were exclusively detected in TPM responders. Of them, 32 genes had nominal P values < 5 × 10^-3 at either week 8 or week 12 and false discovery rates < 0.15 at both time points with consistent directions of gene expression changes, which include GABARAPL1, GPR155, and IL15RA in GABA receptor signaling that represent direct targets for TPM. Analyses of these 300 genes revealed 7 enriched pathways belonging to neuronal function/synaptic plasticity, signal transduction, inflammation/immune function, and oxidative stress response categories. No pathways were enriched for 72 genes exclusively detected in both week 8 and week 12 placebo groups.

CONCLUSION

This secondary analysis study of gene expression data from a TPM clinical trial not only yielded consistent results with those of primary analysis but also identified additional new genes and pathways on TPM response to METH addiction.

Quinacrine overcomes resistance to erlotinib by inhibiting FACT, NF-κB, and cell-cycle progression in non-small cell lung cancer

Erlotinib is a tyrosine kinase inhibitor approved for the treatment of patients with advanced non-small cell lung cancer (NSCLC). In these patients, erlotinib prolongs survival but its benefit remains modest because many tumors express wild-type (wt) EGFR or develop a second-site EGFR mutation. To test drug combinations that could improve the efficacy of erlotinib, we combined erlotinib with quinacrine, which inhibits the FACT (facilitates chromatin transcription) complex that is required for NF-κB transcriptional activity. In A549 (wtEGFR), H1975 (EGFR-L858R/T790M), and H1993 (MET amplification) NSCLC cells, this drug combination was highly synergistic, as quantified by Chou-Talalay combination indices, and slowed xenograft tumor growth. At a sub-IC50 but more clinically attainable concentration of erlotinib, quinacrine, alone or in combination with erlotinib, significantly inhibited colony formation and induced cell-cycle arrest and apoptosis. Quinacrine decreased the level of active FACT subunit SSRP1 and suppressed NF-κB-dependent luciferase activity. Knockdown of SSRP1 decreased cell growth and sensitized cells to erlotinib. Moreover, transcriptomic profiling showed that quinacrine or combination treatment significantly affected cell-cycle-related genes that contain binding sites for transcription factors that regulate SSRP1 target genes. As potential biomarkers of drug combination efficacy, we identified genes that were more strongly suppressed by the combination than by single treatment, and whose increased expression predicted poorer survival in patients with lung adenocarcinoma. This preclinical study shows that quinacrine overcomes erlotinib resistance by inhibiting FACT and cell-cycle progression, and supports a clinical trial testing erlotinib alone versus this combination in advanced NSCLC.

Shifts in the vascular endothelial growth factor isoforms result in transcriptome changes correlated with early neural stem cell proliferation and differentiation in mouse forebrain

Regulation of neural stem cell (NSC) fate decisions is critical during the transition from a multicellular mammalian forebrain neuroepithelium to the multilayered neocortex. Forebrain development requires coordinated vascular investment alongside NSC differentiation. Vascular endothelial growth factor A (Vegf) has proven to be a pleiotrophic gene whose multiple protein isoforms regulate a broad range of effects in neurovascular systems. To test the hypothesis that the Vegf isoforms (120, 164, and 188) are required for normal forebrain development, we analyzed the forebrain transcriptome of mice expressing specific Vegf isoforms, Vegf120, VegfF188, or a combination of Vegf120/188. Transcriptome analysis identified differentially expressed genes in embryonic day (E) 9.5 forebrain, a time point preceding dramatic neuroepithelial expansion and vascular investment in the telencephalon. Meta-analysis identified gene pathways linked to chromosome-level modifications, cell fate regulation, and neurogenesis that were altered in Vegf isoform mice. Based on these gene network shifts, we predicted that NSC populations would be affected in later stages of forebrain development. In the E11.5 telencephalon, we quantified mitotic cells [Phospho-Histone H3 (pHH3)-positive] and intermediate progenitor cells (Tbr2/Eomes-positive), observing quantitative and qualitative shifts in these populations. We observed qualitative shifts in cortical layering at P0, particularly with Ctip2-positive cells in layer V. The results identify a suite of genes and functional gene networks that can be used to further dissect the role of Vegf in regulating NSC differentiation and downstream consequences for NSC fate decisions.

Transcriptome profiling of the newborn mouse lung after hypoxia and reoxygenation: hyperoxic reoxygenation affects mTOR signaling pathway, DNA repair, and JNK-pathway regulation

BACKGROUND

The use of oxygen in acute treatment of asphyxiated term newborns is associated with increased mortality. It is unclear how hyperoxic reoxygenation after hypoxia affects transcriptional changes in the newborn lung.

METHODS

On postnatal day 7, C57BL/6 mice (n = 62) were randomized to 120-min hypoxia (fraction of inspired oxygen (FiO2) 0.08) or normoxia. The hypoxia group was further randomized to reoxygenation for 30 min with FiO2 0.21, 0.40, 0.60, or 1.00, and the normoxia group to FiO2 0.21 or 1.00. Transcriptome profiling was performed on homogenized lung tissue using the Affymetrix 750k expression array, and validation was carried out by real-time polymerase chain reaction and enzyme-linked immunosorbent assay.

RESULTS

The hypoxia-reoxygenation model induced hypoxia-inducible factor 1 (HIF-1) targets like Vegfc, Adm, and Aqp1. In total, ~70% of the significantly differentially expressed genes were detected in the two high hyperoxic groups (FiO2 0.60 and 1.00). Reoxygenation with 100% oxygen after hypoxia uniquely upregulated Gadd45g, Dusp1, Peg3, and Tgm2. Pathway analysis identified mammalian target of rapamycin (mTOR) signaling pathway, DNA repair, c-jun N-terminal kinase (JNK)-pathway regulation, and cell cycle after hyperoxic reoxygenation was applied.

CONCLUSION

Acute hypoxia induces HIF-1 targets independent of the reoxygenation regime applied. Hyperoxic reoxygenation affects pathways regulating cell growth and survival. DNA-damage-responsive genes are restricted to reoxygenation with 100% oxygen.

Literature aided determination of data quality and statistical significance threshold for gene expression studies

Background

Gene expression data are noisy due to technical and biological variability. Consequently, analysis of gene expression data is complex. Different statistical methods produce distinct sets of genes. In addition, selection of expression p-value (EPv) threshold is somewhat arbitrary. In this study, we aimed to develop novel literature based approaches to integrate functional information in analysis of gene expression data.

Methods

Functional relationships between genes were derived by Latent Semantic Indexing (LSI) of Medline abstracts and used to calculate the function cohesion of gene sets. In this study, literature cohesion was applied in two ways. First, Literature-Based Functional Significance (LBFS) method was developed to calculate a p-value for the cohesion of differentially expressed genes (DEGs) in order to objectively evaluate the overall biological significance of the gene expression experiments. Second, Literature Aided Statistical Significance Threshold (LASST) was developed to determine the appropriate expression p-value threshold for a given experiment.

Results

We tested our methods on three different publicly available datasets. LBFS analysis demonstrated that only two experiments were significantly cohesive. For each experiment, we also compared the LBFS values of DEGs generated by four different statistical methods. We found that some statistical tests produced more functionally cohesive gene sets than others. However, no statistical test was consistently better for all experiments. This reemphasizes that a statistical test must be carefully selected for each expression study. Moreover, LASST analysis demonstrated that the expression p-value thresholds for some experiments were considerably lower (p < 0.02 and 0.01), suggesting that the arbitrary p-values and false discovery rate thresholds that are commonly used in expression studies may not be biologically sound.

Conclusions

We have developed robust and objective literature-based methods to evaluate the biological support for gene expression experiments and to determine the appropriate statistical significance threshold. These methods will assist investigators to more efficiently extract biologically meaningful insights from high throughput gene expression experiments.

Stress associated gene expression in blood cells is related to outcome in radiotherapy treated head and neck cancer patients

Background

We previously observed that a radiotherapy-induced biochemical response in plasma was associated with favourable outcome in head and neck squamous carcinoma cancer (HNSCC) patients. The aim of the present study was to compare stress associated blood cell gene expression between two sub-groups of HNSCC patients with different biochemical responses to radiotherapy.

Methods

Out of 87 patients (histologically verified), 10 biochemical ‘responders’ having a high relative increase in plasma oxidative damage and a concomitant decrease in plasma antioxidants during radiotherapy and 10 ‘poor-responders’ were selected for gene-expression analysis and compared using gene set enrichment analysis.

Results

There was a significant induction of stress-relevant gene-sets in the responders following radiotherapy compared to the poor-responders. The relevance of the involvement of similar stress associated gene expression for HNSCC cancer and radioresistance was verified using two publicly available data sets of 42 HNSCC cases and 14 controls (GEO GSE6791), and radiation resistant and radiation sensitive HNSCC xenografts (E-GEOD-9716).

Conclusions

Radiotherapy induces a systemic stress response, as revealed by induction of stress relevant gene expression in blood cells, which is associated to favourable outcome in a cohort of 87 HNSCC patients. Whether these changes in gene expression reflects a systemic effect or are biomarkers of the tumour micro-environmental status needs further study.

Trial registration

Raw data are available at ArrayExpress under accession number E-MEXP-2460.

Identification of novel mouse genes conferring posthypoxic pauses

Although central to the susceptibility of adult diseases characterized by abnormal rhythmogenesis, characterizing the genes involved is a challenge. We took advantage of the C57BL/6J (B6) trait of hypoxia-induced periodic breathing and its absence in the C57BL/6J-Chr 1(A/J)/NaJ chromosome substitution strain to test the feasibility of gene discovery for this abnormality. Beginning with a genetic and phenotypic analysis of an intercross study between these strains, we discovered three quantitative trait loci (QTLs) on mouse chromosome 1, with phenotypic effects. Fine-mapping reduced the genomic intervals and gene content, and the introgression of one QTL region back onto the C57BL/6J-Chr 1(A/J)/NaJ restored the trait. mRNA expression of non-synonymous genes in the introgressed region in the medulla and pons found evidence for differential expression of three genes, the highest of which was apolipoprotein A2, a lipase regulator; the apo a2 peptide fragment (THEQLTPLVR), highly expressed in the liver, was expressed in low amounts in the medulla but did not correlate with trait expression. This work directly demonstrates the impact of elements on mouse chromosome 1 in respiratory rhythmogenesis.

Cyber-T web server: differential analysis of high-throughput data

The Bayesian regularization method for high-throughput differential analysis, described in Baldi and Long (A Bayesian framework for the analysis of microarray expression data: regularized t-test and statistical inferences of gene changes. Bioinformatics 2001: 17: 509-519) and implemented in the Cyber-T web server, is one of the most widely validated. Cyber-T implements a t-test using a Bayesian framework to compute a regularized variance of the measurements associated with each probe under each condition. This regularized estimate is derived by flexibly combining the empirical measurements with a prior, or background, derived from pooling measurements associated with probes in the same neighborhood. This approach flexibly addresses problems associated with low replication levels and technology biases, not only for DNA microarrays, but also for other technologies, such as protein arrays, quantitative mass spectrometry and next-generation sequencing (RNA-seq). Here we present an update to the Cyber-T web server, incorporating several useful new additions and improvements. Several preprocessing data normalization options including logarithmic and (Variance Stabilizing Normalization) VSN transforms are included. To augment two-sample t-tests, a one-way analysis of variance is implemented. Several methods for multiple tests correction, including standard frequentist methods and a probabilistic mixture model treatment, are available. Diagnostic plots allow visual assessment of the results. The web server provides comprehensive documentation and example data sets. The Cyber-T web server, with R source code and data sets, is publicly available at http://cybert.ics.uci.edu/.

Mapping the interaction of Snf1 with TORC1 in Saccharomyces cerevisiae

Nutrient sensing and coordination of metabolic pathways are crucial functions for living cells. A combined analysis of the yeast transcriptome, phosphoproteome and metabolome is used to investigate the interactions between the Snf1 and TORC1 pathways under nutrient-limited conditions.

Snf1 regulates a broad range of biological processes, while target of rapamycin complex 1 (TORC1) seems to be repressed under both glucose- and ammonium-limited conditions.

Snf1 has a role in regulating amino acids by upregulating the NADP+-dependent glutamate dehydrogenase (encoded by GDH3) under glucose-limited condition.

In addition to the accepted role of Snf1 in regulating fatty acid (FA) metabolism, TORC1 may also regulate FA metabolism.

Direct interactions between Snf1 and TORC1 pathways are unlikely under nutrient-limited conditions and TORC1 might be repressed in a manner that is independent of Snf1.

Nutrient sensing and coordination of metabolic pathways are crucial functions for all living cells, but details of the coordination under different environmental conditions remain elusive. We therefore undertook a systems biology approach to investigate the interactions between the Snf1 and the target of rapamycin complex 1 (TORC1) in Saccharomyces cerevisiae. We show that Snf1 regulates a much broader range of biological processes compared with TORC1 under both glucose- and ammonium-limited conditions. We also find that Snf1 has a role in upregulating the NADP⁺-dependent glutamate dehydrogenase (encoded by GDH3) under derepressing condition, and therefore may also have a role in ammonium assimilation and amino-acid biosynthesis, which can be considered as a convergence of Snf1 and TORC1 pathways. In addition to the accepted role of Snf1 in regulating fatty acid (FA) metabolism, we show that TORC1 also regulates FA metabolism, likely through modulating the peroxisome and β-oxidation. Finally, we conclude that direct interactions between Snf1 and TORC1 pathways are unlikely under nutrient-limited conditions and propose that TORC1 is repressed in a manner that is independent of Snf1.

Genome-wide gene expression analysis suggests an important role of hypoxia in the pathogenesis of endemic osteochondropathy Kashin-Beck disease

Kashin-Beck Disease (KBD) is an endemic osteochondropathy, the pathogenesis of which remains unclear now. In this study, we compared gene expression profiles of articular cartilage derived respectively from KBD patients and normal controls. Total RNA were isolated, amplified, labeled and hybridized to Agilent human 1A 22 k whole genome microarray chip. qRT-PCR was conducted to validate our microarray data. We detected 57 up-regulated genes (ratios ≥2.0) and 24 down-regulated genes (ratios ≤0.5) in KBD cartilage. To further identify the key genes involved in the pathogenesis of KBD, Bayesian analysis of variance for microarrays (BAM) software was applied and identified 12 potential key genes with an average ratio 6.64, involved in apoptosis, metabolism, cytokine & growth factor and cytoskeleton & cell movement. Gene Set Enrichment Analysis (GSEA) software was used to identify differently expressed gene ontology categories and pathways. GSEA found that a set of apoptosis, hypoxia and mitochondrial function related gene ontology categories and pathways were significantly up-regulated in KBD compared to normal controls. Based on the results of this study, we suggest that chronic hypoxia-induced mitochondrial damage and apoptosis might play an important role in the pathogenesis of KBD. Our efforts may help to understand the pathogenesis of KBD as well as other osteoarthrosis with similar articular cartilage lesions.

Blood cell gene expression associated with cellular stress defense is modulated by antioxidant-rich food in a randomised controlled clinical trial of male smokers

BACKGROUND

Plant-based diets rich in fruit and vegetables can prevent development of several chronic age-related diseases. However, the mechanisms behind this protective effect are not elucidated. We have tested the hypothesis that intake of antioxidant-rich foods can affect groups of genes associated with cellular stress defence in human blood cells.

TRIAL REGISTRATION NUMBER

NCT00520819 http://clinicaltrials.gov.

METHODS

In an 8-week dietary intervention study, 102 healthy male smokers were randomised to either a diet rich in various antioxidant-rich foods, a kiwifruit diet (three kiwifruits/d added to the regular diet) or a control group. Blood cell gene expression profiles were obtained from 10 randomly selected individuals of each group. Diet-induced changes on gene expression were compared to controls using a novel application of the gene set enrichment analysis (GSEA) on transcription profiles obtained using Affymetrix HG-U133-Plus 2.0 whole genome arrays.

RESULTS

Changes were observed in the blood cell gene expression profiles in both intervention groups when compared to the control group. Groups of genes involved in regulation of cellular stress defence, such as DNA repair, apoptosis and hypoxia, were significantly upregulated (GSEA, FDR q-values < 5%) by both diets compared to the control group. Genes with common regulatory motifs for aryl hydrocarbon receptor (AhR) and AhR nuclear translocator (AhR/ARNT) were upregulated by both interventions (FDR q-values < 5%). Plasma antioxidant biomarkers (polyphenols/carotenoids) increased in both groups.

CONCLUSIONS

The observed changes in the blood cell gene expression profiles suggest that the beneficial effects of a plant-based diet on human health may be mediated through optimization of defence processes.

A mouse embryonic stem cell bank for inducible overexpression of human chromosome 21 genes

Background

Dosage imbalance is responsible for several genetic diseases, among which Down syndrome is caused by the trisomy of human chromosome 21.

Results

To elucidate the extent to which the dosage imbalance of specific human chromosome 21 genes perturb distinct molecular pathways, we developed the first mouse embryonic stem (ES) cell bank of human chromosome 21 genes. The human chromosome 21-mouse ES cell bank includes, in triplicate clones, 32 human chromosome 21 genes, which can be overexpressed in an inducible manner. Each clone was transcriptionally profiled in inducing versus non-inducing conditions. Analysis of the transcriptional response yielded results that were consistent with the perturbed gene's known function. Comparison between mouse ES cells containing the whole human chromosome 21 (trisomic mouse ES cells) and mouse ES cells overexpressing single human chromosome 21 genes allowed us to evaluate the contribution of single genes to the trisomic mouse ES cell transcriptome. In addition, for the clones overexpressing the Runx1 gene, we compared the transcriptome changes with the corresponding protein changes by mass spectroscopy analysis.

Conclusions

We determined that only a subset of genes produces a strong transcriptional response when overexpressed in mouse ES cells and that this effect can be predicted taking into account the basal gene expression level and the protein secondary structure. We showed that the human chromosome 21-mouse ES cell bank is an important resource, which may be instrumental towards a better understanding of Down syndrome and other human aneuploidy disorders.

Gene signature is associated with early stage rectal cancer recurrence

BACKGROUND

Despite expected excellent outcomes of surgical resection for early stage rectal cancers, 20% of stage I and II rectal cancers recur. Identifying biologic factors that predict the subset prone to recur could allow more directed therapy. This study identifies a tumor gene expression profile that accurately predicts disease recurrence.

STUDY DESIGN

Stage I/II rectal cancer patients treated by surgery alone at a single institution were included and classified as having recurrent or nonrecurrent cancer. Tumor mRNA was isolated from frozen tissue and evaluated for total genome gene expression by microarray analysis. Background-corrected and normalized microarray data were analyzed using BAMarray software. Selected genes were further analyzed using unsupervised clustering and nearest-centroid classification. A balanced K-fold scoring-pair algorithm using 1,000 independent replications was used for gene signature development.

RESULTS

Sixty-nine patients with disease-free survival and 31 patients with recurrent disease were included at a median follow-up of 105 months (interquartile range 114 months) and 32 months (interquartile range 25 months), respectively. Demographics and tumor characteristics between groups were similar. Fifty-two genes from 43,148 probes were differentially expressed, and a 36-gene signature was found to be statistically associated with recurrence using a scoring-pair algorithm. Accuracy to identify recurrence as measured by area under the receiver operating characteristic curve was 0.803.

CONCLUSIONS

Differential gene expression within rectal cancers is associated with recurrence of early stage disease. A 36-gene signature correlates with an increased risk of more or less aggressive tumor behavior. This information obtainable at biopsy may assist in determining treatment decisions.

Challenges and approaches to statistical design and inference in high-dimensional investigations

Advances in modern technologies have facilitated high-dimensional experiments (HDEs) that generate tremendous amounts of genomic, proteomic, and other "omic" data. HDEs involving whole-genome sequences and polymorphisms, expression levels of genes, protein abundance measurements, and combinations thereof have become a vanguard for new analytic approaches to the analysis of HDE data. Such situations demand creative approaches to the processes of statistical inference, estimation, prediction, classification, and study design. The novel and challenging biological questions asked from HDE data have resulted in many specialized analytic techniques being developed. This chapter discusses some of the unique statistical challenges facing investigators studying high-dimensional biology and describes some approaches being developed by statistical scientists. We have included some focus on the increasing interest in questions involving testing multiple propositions simultaneously, appropriate inferential indicators for the types of questions biologists are interested in, and the need for replication of results across independent studies, investigators, and settings. A key consideration inherent throughout is the challenge in providing methods that a statistician judges to be sound and a biologist finds informative.

Genomic promoter analysis predicts functional transcription factor binding

Background. The computational identification of functional transcription factor binding sites (TFBSs) remains a major challenge of computational biology. Results. We have analyzed the conserved promoter sequences for the complete set of human RefSeq genes using our conserved transcription factor binding site (CONFAC) software. CONFAC identified 16296 human-mouse ortholog gene pairs, and of those pairs, 9107 genes contained conserved TFBS in the 3 kb proximal promoter and first intron. To attempt to predict in vivo occupancy of transcription factor binding sites, we developed a novel marginal effect isolator algorithm that builds upon Bayesian methods for multigroup TFBS filtering and predicted the in vivo occupancy of two transcription factors with an overall accuracy of 84%. Conclusion. Our analyses show that integration of chromatin immunoprecipitation data with conserved TFBS analysis can be used to generate accurate predictions of functional TFBS. They also show that TFBS cooccurrence can be used to predict transcription factor binding to promoters in vivo.

Studying genetic determinants of natural variation in human gene expression using Bayesian ANOVA

Standard genetic mapping techniques scan chromosomal segments for location of genetic linkage and association signals. The majority of these methods consider only correlations at single markers and/or phenotypes with explicit detailing of the genetic structure. These methods tend to be limited by their inability to consider the effect of large numbers of model variables jointly. In contrast, we propose a Bayesian analysis of variance (ANOVA) method to categorize individuals based on similarity of multidimensional profiles and attempt to analyze all variables simultaneously. Using Problem 1 of the Genetic Analysis Workshop 15 data set, we demonstrate the method's utility for joint analysis of gene expression levels and single-nucleotide polymorphism genotypes. We show that the method extracts similar information to that of previous genetic mapping analyses, and suggest extensions of the method for mining unique information not previously found.

From trees to wood and back: perspective on clinical data analysis in thoracic surgery

General thoracic surgery has been slow to embrace advanced statistical techniques to assess outcome and direct future therapy to individual patients (the "trees") based on aggregating data and discovering informative patterns in them (the "wood"). This is due in part to the nature and complexity of the practice. However, methods exist that deal with problems of small numbers, variable case volume, and heterogeneity of patients, diseases, and therapies. For effective analysis of these multiple variables, modern strategies must be used and pitfalls avoided. For generating new knowledge about appropriateness of therapy, the focus of analyses must include long-term outcome. Quality improvement efforts may be assisted by in-depth analysis of short-term outcome from a select number of centers to inform future universal efforts. Ultimately, these results need to feed back to improved management of the individual patient.

Oxidoreductase, morphogenesis, extracellular matrix, and calcium ion-binding gene expression in streptozotocin-induced diabetic rat heart

Diabetes has far-ranging effects on cardiac structure and function. Previous gene expression studies of the heart in animal models of type 1 diabetes concur that there is altered expression of genes involved in lipid and protein metabolism, but they diverge with regard to expression changes involving many other functional groups of genes of mechanistic importance in diabetes-induced cardiac dysfunction. To obtain additional information about these controversial areas, genome-wide expression was assessed using microarrays in left ventricle from streptozotocin-diabetic and normal rats. There were 261 genes with statistically significant altered expression of at least +/-1.5-fold, of which 124 were increased and 137 reduced by diabetes. Gene ontology assignment testing identified several statistical significantly overrepresented groups among genes with altered expression, which differed for increased compared with reduced expression. Relevant gene groups with increased expression by diabetes included lipid metabolism (P < 0.001, n = 13 genes, fold change 1.5 to 14.6) and oxidoreductase activity (P < 0.001, n = 17, fold change 1.5 to 4.6). Groups with reduced expression by diabetes included morphogenesis (P < 0.00001, n = 28, fold change -1.5 to -5.1), extracellular matrix (P < 0.02, n = 9, fold change -1.5 to -3.9), cell adhesion (P < 0.05, n = 10, fold change -1.5 to -2.7), and calcium ion binding (P < 0.01, n = 13, fold change -1.5 to -3.0). Array findings were verified by quantitative PCR for 36 genes. These data combined with previous findings strengthen the evidence for diabetes-induced cardiac gene expression changes involved in cell growth and development, oxidoreductase activity, and the extracellular matrix and also point out other gene groups not previously identified as being affected, such as those involved in calcium ion homeostasis.

Epistemological issues in omics and high-dimensional biology: give the people what they want

Gene expression microarrays have been the vanguard of new analytic approaches in high-dimensional biology. Draft sequences of several genomes coupled with new technologies allow study of the influences and responses of entire genomes rather than isolated genes. This has opened a new realm of highly dimensional biology where questions involve multiplicity at unprecedented scales: thousands of genetic polymorphisms, gene expression levels, protein measurements, genetic sequences, or any combination of these and their interactions. Such situations demand creative approaches to the processes of inference, estimation, prediction, classification, and study design. Although bench scientists intuitively grasp the need for flexibility in the inferential process, the elaboration of formal supporting statistical frameworks is just at the very start. Here, we will discuss some of the unique statistical challenges facing investigators studying high-dimensional biology, describe some approaches being developed by statistical scientists, and offer an epistemological framework for the validation of proffered statistical procedures. A key theme will be the challenge in providing methods that a statistician judges to be sound and a biologist finds informative. The shift from family-wise error rate control to false discovery rate estimation and to assessment of ranking and other forms of stability will be portrayed as illustrative of approaches to this challenge.

Advances in statistical human genetics over the last 25 years

The past 25 years has seen an explosion in the number of genetic markers that can be measured on DNA samples at an ever decreasing cost. Although basic statistical methods for analysing such data gathered on samples of either independent individuals or family members, one or two markers at a time, were already well developed before this explosion occurred, there has been a corresponding burst in activity to develop multiple marker models to find disease-causing gene variants, capitalizing on the data that have become available, to increase the power of such methods. This has required the concomitant development of faster algorithms to speed up the computation of various likelihoods. For linkage analysis, to obtain the approximate locations for genes of interest, Mendelian segregation models have been extended to be more realistic and statistical models that do not assume specific modes of inheritance have been extended to allow for the analysis of larger pedigree structures. For association analysis, to obtain more precise locations for genes of interest, the recent completion of the first stage of the HapMap project has spurred the development, still underway, of novel experimental designs and analytical methods to combat the curse of dimensionality and the resulting multiple testing problem. Perhaps the greatest current challenge concerns how best to gather and synthesize the many lines of evidence possible in order to discover the genetic determinants underlying complex diseases.