Rat toxicogenomic study reveals analytical consistency across microarray platforms

The balance of reproducibility, sensitivity, and specificity of lists of differentially expressed genes in microarray studies

BACKGROUND

Reproducibility is a fundamental requirement in scientific experiments. Some recent publications have claimed that microarrays are unreliable because lists of differentially expressed genes (DEGs) are not reproducible in similar experiments. Meanwhile, new statistical methods for identifying DEGs continue to appear in the scientific literature. The resultant variety of existing and emerging methods exacerbates confusion and continuing debate in the microarray community on the appropriate choice of methods for identifying reliable DEG lists.

RESULTS

Using the data sets generated by the MicroArray Quality Control (MAQC) project, we investigated the impact on the reproducibility of DEG lists of a few widely used gene selection procedures. We present comprehensive results from inter-site comparisons using the same microarray platform, cross-platform comparisons using multiple microarray platforms, and comparisons between microarray results and those from TaqMan - the widely regarded "standard" gene expression platform. Our results demonstrate that (1) previously reported discordance between DEG lists could simply result from ranking and selecting DEGs solely by statistical significance (P) derived from widely used simple t-tests; (2) when fold change (FC) is used as the ranking criterion with a non-stringent P-value cutoff filtering, the DEG lists become much more reproducible, especially when fewer genes are selected as differentially expressed, as is the case in most microarray studies; and (3) the instability of short DEG lists solely based on P-value ranking is an expected mathematical consequence of the high variability of the t-values; the more stringent the P-value threshold, the less reproducible the DEG list is. These observations are also consistent with results from extensive simulation calculations.

CONCLUSION

We recommend the use of FC-ranking plus a non-stringent P cutoff as a straightforward and baseline practice in order to generate more reproducible DEG lists. Specifically, the P-value cutoff should not be stringent (too small) and FC should be as large as possible. Our results provide practical guidance to choose the appropriate FC and P-value cutoffs when selecting a given number of DEGs. The FC criterion enhances reproducibility, whereas the P criterion balances sensitivity and specificity.

Identification of candidate predictive and surrogate molecular markers for dasatinib in prostate cancer: rationale for patient selection and efficacy monitoring

BACKGROUND

Dasatinib is a potent, multi-targeted kinase inhibitor that was recently approved for treatment of chronic myelogenous leukemia resistant to imatinib. To aid the clinical development of dasatinib in prostate cancer, we utilized preclinical models to identify potential molecular markers for patient stratification and efficacy monitoring.

RESULTS

Using gene expression profiling, we first identified 174 genes whose expression was highly correlated with in vitro sensitivity of 16 cell lines and, thus, considered as candidate efficacy predictive markers. Among these are important prostatic cell lineage markers, cytokeratin 5, androgen receptor and prostate specific antigen. Our results indicate that 'basal type' cell lines with high expression of cytokeratin 5 and low expression of androgen receptor or prostate specific antigen are sensitive to dasatinib. To identify markers as surrogates for biological activity, we treated cell lines with dasatinib and identified genes whose expression was significantly modulated by the drug. Ten genes, including that encoding urokinase-type plasminogen activator (uPA), were found to not only be potential efficacy markers but also to have reduced expression upon dasatinib treatment. The down-regulation of uPA by dasatinib was drug-specific and correlated with the sensitivity of cell lines to dasatinib. Furthermore, EphA2, a target of dasatinib, was found to be a sensitivity biomarker.

CONCLUSION

Using the gene expression profiling approach and preclinical models, we have identified prostatic biomarkers that are associated with sensitivity to dasatinib. This study has provided a basis for clinical evaluation of a potential dasatinib efficacy signature in prostate cancer.

Thyroid hormone action in the adult brain: gene expression profiling of the effects of single and multiple doses of triiodo-L-thyronine in the rat striatum

Thyroid hormones have profound effects on mood and behavior, but the molecular basis of thyroid hormone action in the adult brain is relatively unknown. In particular, few thyroid hormone-dependent genes have been identified in the adult brain despite extensive work carried out on the developing brain. In this work we performed global analysis of gene expression in the adult rat striatum in search for genomic changes taking place after administration of T(3) to hypothyroid rats. The hormone was administered in two different schedules: 1) a single, large dose of 25 microg per 100 g body weight (SD) or 2) 1.5 microg per 100 g body weight once daily for 5 d (RD). Twenty-four hours after the single or last of multiple doses, gene expression in the striatum was analyzed using Codelink microarrays. SD caused up-regulation of 149 genes and down-regulation of 88 genes. RD caused up-regulation of 18 genes and down-regulation of one gene. The results were confirmed by hybridization to Affymetrix microarrays and by TaqMan PCR. Among the genes identified are genes involved in circadian regulation and the regulation of signaling pathways in the striatum. These results suggest that thyroid hormone is involved in regulation of striatal physiology at multiple control points. In addition, they may explain the beneficial effects of large doses of thyroid hormone in bipolar disorders.

Apparently low reproducibility of true differential expression discoveries in microarray studies

MOTIVATION

Differentially expressed gene (DEG) lists detected from different microarray studies for a same disease are often highly inconsistent. Even in technical replicate tests using identical samples, DEG detection still shows very low reproducibility. It is often believed that current small microarray studies will largely introduce false discoveries.

RESULTS

Based on a statistical model, we show that even in technical replicate tests using identical samples, it is highly likely that the selected DEG lists will be very inconsistent in the presence of small measurement variations. Therefore, the apparently low reproducibility of DEG detection from current technical replicate tests does not indicate low quality of microarray technology. We also demonstrate that heterogeneous biological variations existing in real cancer data will further reduce the overall reproducibility of DEG detection. Nevertheless, in small subsamples from both simulated and real data, the actual false discovery rate (FDR) for each DEG list tends to be low, suggesting that each separately determined list may comprise mostly true DEGs. Rather than simply counting the overlaps of the discovery lists from different studies for a complex disease, novel metrics are needed for evaluating the reproducibility of discoveries characterized with correlated molecular changes. Supplementaty information: Supplementary data are available at Bioinformatics online.

Efficiency analysis of competing tests for finding differentially expressed genes in lung adenocarcinoma

In this study, we introduce and use Efficiency Analysis to compare differences in the apparent internal and external consistency of competing normalization methods and tests for identifying differentially expressed genes. Using publicly available data, two lung adenocarcinoma datasets were analyzed using caGEDA (http://bioinformatics2.pitt.edu/GE2/GEDA.html) to measure the degree of differential expression of genes existing between two populations. The datasets were randomly split into at least two subsets, each analyzed for differentially expressed genes between the two sample groups, and the gene lists compared for overlapping genes. Efficiency Analysis is an intuitive method that compares the differences in the percentage of overlap of genes from two or more data subsets, found by the same test over a range of testing methods. Tests that yield consistent gene lists across independently analyzed splits are preferred to those that yield less consistent inferences. For example, a method that exhibits 50% overlap in the 100 top genes from two studies should be preferred to a method that exhibits 5% overlap in the top 100 genes. The same procedure was performed using all available normalization and transformation methods that are available through caGEDA. The ‘best’ test was then further evaluated using internal cross-validation to estimate generalizable sample classification errors using a Naïve Bayes classification algorithm. A novel test, termed D1 (a derivative of the J5 test) was found to be the most consistent, and to exhibit the lowest overall classification error, and highest sensitivity and specificity. The D1 test relaxes the assumption that few genes are differentially expressed. Efficiency Analysis can be misleading if the tests exhibit a bias in any particular dimension (e.g. expression intensity); we therefore explored intensity-scaled and segmented J5 tests using data in which all genes are scaled to share the same intensity distribution range. Efficiency Analysis correctly predicted the ‘best’ test and normalization method using the Beer dataset and also performed well with the Bhattacharjee dataset based on both efficiency and classification accuracy criteria.

Cross-platform comparison of SYBR Green real-time PCR with TaqMan PCR, microarrays and other gene expression measurement technologies evaluated in the MicroArray Quality Control (MAQC) study

Background

The MicroArray Quality Control (MAQC) project evaluated the inter- and intra-platform reproducibility of seven microarray platforms and three quantitative gene expression assays in profiling the expression of two commercially available Reference RNA samples (Nat Biotechnol 24:1115-22, 2006). The tested microarrays were the platforms from Affymetrix, Agilent Technologies, Applied Biosystems, GE Healthcare, Illumina, Eppendorf and the National Cancer Institute, and quantitative gene expression assays included TaqMan^® Gene Expression PCR Assay, Standardized (Sta) RT-PCR™ and QuantiGene^®. The data showed great consistency in gene expression measurements across different microarray platforms, different technologies and test sites. However, SYBR^® Green real-time PCR, another common technique utilized by half of all real-time PCR users for gene expression measurement, was not addressed in the MAQC study. In the present study, we compared the performance of SYBR Green PCR with TaqMan PCR, microarrays and other quantitative technologies using the same two Reference RNA samples as the MAQC project. We assessed SYBR Green real-time PCR using commercially available RT² Profiler™ PCR Arrays from SuperArray, containing primer pairs that have been experimentally validated to ensure gene-specificity and high amplification efficiency.

Results

The SYBR Green PCR Arrays exhibit good reproducibility among different users, PCR instruments and test sites. In addition, the SYBR Green PCR Arrays have the highest concordance with TaqMan PCR, and a high level of concordance with other quantitative methods and microarrays that were evaluated in this study in terms of fold-change correlation and overlap of lists of differentially expressed genes.

Conclusion

These data demonstrate that SYBR Green real-time PCR delivers highly comparable results in gene expression measurement with TaqMan PCR and other high-density microarrays.

Performance comparison of two microarray platforms to assess differential gene expression in human monocyte and macrophage cells

Background

In this study we assessed the respective ability of Affymetrix and Illumina microarray methodologies to answer a relevant biological question, namely the change in gene expression between resting monocytes and macrophages derived from these monocytes. Five RNA samples for each type of cell were hybridized to the two platforms in parallel. In addition, a reference list of differentially expressed genes (DEG) was generated from a larger number of hybridizations (mRNA from 86 individuals) using the RNG/MRC two-color platform.

Results

Our results show an important overlap of the Illumina and Affymetrix DEG lists. In addition, more than 70% of the genes in these lists were also present in the reference list. Overall the two platforms had very similar performance in terms of biological significance, evaluated by the presence in the DEG lists of an excess of genes belonging to Gene Ontology (GO) categories relevant for the biology of monocytes and macrophages. Our results support the conclusion of the MicroArray Quality Control (MAQC) project that the criteria used to constitute the DEG lists strongly influence the degree of concordance among platforms. However the importance of prioritizing genes by magnitude of effect (fold change) rather than statistical significance (p-value) to enhance cross-platform reproducibility recommended by the MAQC authors was not supported by our data.

Conclusion

Functional analysis based on GO enrichment demonstrates that the 2 compared technologies delivered very similar results and identified most of the relevant GO categories enriched in the reference list.

Gene expression response in target organ and whole blood varies as a function of target organ injury phenotype

This report details the standardized experimental design and the different data streams that were collected (histopathology, clinical chemistry, hematology and gene expression from the target tissue (liver) and a bio-available tissue (blood)) after treatment with eight known hepatotoxicants (at multiple time points and doses with multiple biological replicates). The results of the study demonstrate the classification of histopathological differences, likely reflecting differences in mechanisms of cell-specific toxicity, using either liver tissue or blood transcriptomic data.

Distinct roles of adenylyl cyclases 1 and 8 in opiate dependence: behavioral, electrophysiological, and molecular studies

BACKGROUND

Opiate dependence is a result of adaptive changes in signal transduction networks in several brain regions. Noradrenergic neurons of the locus coeruleus (LC) have provided a useful model system in which to understand the molecular basis of these adaptive changes. One of most robust signaling adaptations to repeated morphine exposure in this brain region is upregulation of adenylyl cyclase (AC) activity. Earlier work revealed the selective induction of two calmodulin-dependent AC isoforms, AC1 and AC8, after chronic morphine, but their role in opiate dependence has remained unknown.

METHODS

Whole cell recordings from LC slices, behavioral paradigms for dependence, and gene array technology have been used to dissect the role of AC1 and AC8 in chronic morphine responses.

RESULTS

Both AC1 and AC8 knockout mice exhibit reduced opiate dependence on the basis of attenuated withdrawal; however, partially distinct withdrawal symptoms were affected in the two lines. Loss of AC1 or AC8 also attenuated the electrophysiological effects of morphine on LC neurons: knockout of either cyclase attenuated the chronic morphine-induced enhancement of baseline firing rates as well as of regulation of neuronal firing by forskolin (an activator of ACs). The DNA microarray analysis revealed that both AC1 and AC8 affect gene regulation in the LC by chronic morphine and, in addition to common genes, each cyclase influences the expression of a distinct subset of genes.

CONCLUSIONS

Together, these findings provide fundamentally new insight into the molecular and cellular basis of opiate dependence.

Improving the power for detecting overlapping genes from multiple DNA microarray-derived gene lists

Background

In DNA microarray gene expression profiling studies, a fundamental task is to extract statistically significant genes that meet certain research hypothesis. Currently, Venn diagram is a frequently used method for identifying overlapping genes that meet the investigator's research hypotheses. However this simple operation of intersecting multiple gene lists, known as the Intersection-Union Tests (IUTs), is performed without knowing the incurred changes in Type 1 error rate and can lead to loss of discovery power.

Results

We developed an IUT adjustment procedure, called Relaxed IUT (RIUT), which is proved to be less conservative and more powerful for intersecting independent tests than the traditional Venn diagram approach. The advantage of the RIUT procedure over traditional IUT is demonstrated by empirical Monte-Carlo simulation and two real toxicogenomic gene expression case studies. Notably, the enhanced power of RIUT enables it to identify overlapping gene sets leading to identification of certain known related pathways which were not detected using the traditional IUT method.

Conclusion

We showed that traditional IUT via a Venn diagram is generally conservative, which may lead to loss discovery power in DNA microarray studies. RIUT is proved to be a more powerful alternative for performing IUTs in identifying overlapping genes from multiple gene lists derived from microarray gene expression profiling.

Enhanced identification and biological validation of differential gene expression via Illumina whole-genome expression arrays through the use of the model-based background correction methodology

Despite the tremendous growth of microarray usage in scientific studies, there is a lack of standards for background correction methodologies, especially in single-color microarray platforms. Traditional background subtraction methods often generate negative signals and thus cause large amounts of data loss. Hence, some researchers prefer to avoid background corrections, which typically result in the underestimation of differential expression. Here, by utilizing nonspecific negative control features integrated into Illumina whole genome expression arrays, we have developed a method of model-based background correction for BeadArrays (MBCB). We compared the MBCB with a method adapted from the Affymetrix robust multi-array analysis algorithm and with no background subtraction, using a mouse acute myeloid leukemia (AML) dataset. We demonstrated that differential expression ratios obtained by using the MBCB had the best correlation with quantitative RT–PCR. MBCB also achieved better sensitivity in detecting differentially expressed genes with biological significance. For example, we demonstrated that the differential regulation of Tnfr2, Ikk and NF-kappaB, the death receptor pathway, in the AML samples, could only be detected by using data after MBCB implementation. We conclude that MBCB is a robust background correction method that will lead to more precise determination of gene expression and better biological interpretation of Illumina BeadArray data.

Effect of astrocyte-targeted production of IL-6 on traumatic brain injury and its impact on the cortical transcriptome

Interleukin-6 (IL-6) is one of the key players in the response of the brain cortex to injury. We have described previously that astrocyte-driven production of IL-6 (GFAP-IL6) in transgenic mice, although causing spontaneous neuroinflammation and long term damage, is beneficial after an acute (freeze) injury in the cortex, increasing healing and decreasing oxidative stress and apoptosis. To determine the transcriptional basis for these responses here we analyzed the global gene expression profile of the cortex, at 0 (unlesioned), 1 or 4 days post lesion (dpl), in both GFAP-IL6 mice and their control littermates. GFAP-IL6 mice showed an increase in genes associated with the inflammatory response both at 1 dpl (Iftm1, Endod1) and 4 dpl (Gfap, C4b), decreased expression of proapoptotic genes (i.e. Gadd45b, Clic4, p21) as well as reduced expression of genes involved in the control of oxidative stress (Atf4). Furthermore, the presence of IL-6 altered the expression of genes involved in hemostasis (Vwf), cell migration and proliferation (Cap2), and synaptic activity (Vamp2). All these changes in gene expression could underlie the phenotype of the GFAP-IL6 mice after injury, but many other possible factors were also identified in this study, highlighting the utility of this approach for deciphering new pathways orchestrated by IL-6.

Technical variables in high-throughput miRNA expression profiling: much work remains to be done

MicroRNA (miRNA) gene expression profiling has provided important insights into plant and animal biology. However, there has not been ample published work about pitfalls associated with technical parameters in miRNA gene expression profiling. One source of pertinent information about technical variables in gene expression profiling is the separate and more well-established literature regarding mRNA expression profiling. However, many aspects of miRNA biochemistry are unique. For example, the cellular processing and compartmentation of miRNAs, the differential stability of specific miRNAs, and aspects of global miRNA expression regulation require specific consideration. Additional possible sources of systematic bias in miRNA expression studies include the differential impact of pre-analytical variables, substrate specificity of nucleic acid processing enzymes used in labeling and amplification, and issues regarding new miRNA discovery and annotation. We conclude that greater focus on technical parameters is required to bolster the validity, reliability, and cultural credibility of miRNA gene expression profiling studies.

Characterization of the proneural gene regulatory network during mouse telencephalon development

Background

The proneural proteins Mash1 and Ngn2 are key cell autonomous regulators of neurogenesis in the mammalian central nervous system, yet little is known about the molecular pathways regulated by these transcription factors.

Results

Here we identify the downstream effectors of proneural genes in the telencephalon using a genomic approach to analyze the transcriptome of mice that are either lacking or overexpressing proneural genes. Novel targets of Ngn2 and/or Mash1 were identified, such as members of the Notch and Wnt pathways, and proteins involved in adhesion and signal transduction. Next, we searched the non-coding sequence surrounding the predicted proneural downstream effector genes for evolutionarily conserved transcription factor binding sites associated with newly defined consensus binding sites for Ngn2 and Mash1. This allowed us to identify potential novel co-factors and co-regulators for proneural proteins, including Creb, Tcf/Lef, Pou-domain containing transcription factors, Sox9, and Mef2a. Finally, a gene regulatory network was delineated using a novel Bayesian-based algorithm that can incorporate information from diverse datasets.

Conclusion

Together, these data shed light on the molecular pathways regulated by proneural genes and demonstrate that the integration of experimentation with bioinformatics can guide both hypothesis testing and hypothesis generation.

Pathway analysis tools and toxicogenomics reference databases for risk assessment

The pharmaceutical industry has begun to leverage a range of new technologies (proteomics, pharmacogenomics, metabolomics and molecular toxicology [e.g., toxicogenomics]) and analysis tools that are becoming increasingly integrated in the area of drug discovery and development. The approach of analyzing the vast amount of toxicogenomics data generated using molecular pathway and networks analysis tools in combination with analysis of reference data will be the main focus of this review. We will demonstrate how this combined approach can increase the understanding of the molecular mechanisms that lead to chemical-induced toxicity and application of this knowledge to compound risk assessment. We will provide an example of the insights achieved through a molecular toxicology analysis based on the well-known hepatotoxicant lipopolysaccharide to illustrate the utility of these new tools in the analysis of complex data sets, both in vivo and in vitro. The ultimate objective is a better lead selection process that improves the chances for success across the different stages of drug discovery and development.

The Songbird Neurogenomics (SoNG) Initiative: community-based tools and strategies for study of brain gene function and evolution

Background

Songbirds hold great promise for biomedical, environmental and evolutionary research. A complete draft sequence of the zebra finch genome is imminent, yet a need remains for application of genomic resources within a research community traditionally focused on ethology and neurobiological methods. In response, we developed a core set of genomic tools and a novel collaborative strategy to probe gene expression in diverse songbird species and natural contexts.

Results

We end-sequenced cDNAs from zebra finch brain and incorporated additional sequences from community sources into a database of 86,784 high quality reads. These assembled into 31,658 non-redundant contigs and singletons, which we annotated via BLAST search of chicken and human databases. The results are publicly available in the ESTIMA:Songbird database. We produced a spotted cDNA microarray with 20,160 addresses representing 17,214 non-redundant products of an estimated 11,500–15,000 genes, validating it by analysis of immediate-early gene (zenk) gene activation following song exposure and by demonstrating effective cross hybridization to genomic DNAs of other songbird species in the Passerida Parvorder. Our assembly was also used in the design of the "Lund-zfa" Affymetrix array representing ~22,000 non-redundant sequences. When the two arrays were hybridized to cDNAs from the same set of male and female zebra finch brain samples, both arrays detected a common set of regulated transcripts with a Pearson correlation coefficient of 0.895. To stimulate use of these resources by the songbird research community and to maintain consistent technical standards, we devised a "Community Collaboration" mechanism whereby individual birdsong researchers develop experiments and provide tissues, but a single individual in the community is responsible for all RNA extractions, labelling and microarray hybridizations.

Conclusion

Immediately, these results set the foundation for a coordinated set of 25 planned experiments by 16 research groups probing fundamental links between genome, brain, evolution and behavior in songbirds. Energetic application of genomic resources to research using songbirds should help illuminate how complex neural and behavioral traits emerge and evolve.

Can subtle changes in gene expression be consistently detected with different microarray platforms?

Background

The comparability of gene expression data generated with different microarray platforms is still a matter of concern. Here we address the performance and the overlap in the detection of differentially expressed genes for five different microarray platforms in a challenging biological context where differences in gene expression are few and subtle.

Results

Gene expression profiles in the hippocampus of five wild-type and five transgenic δC-doublecortin-like kinase mice were evaluated with five microarray platforms: Applied Biosystems, Affymetrix, Agilent, Illumina, LGTC home-spotted arrays. Using a fixed false discovery rate of 10% we detected surprising differences between the number of differentially expressed genes per platform. Four genes were selected by ABI, 130 by Affymetrix, 3,051 by Agilent, 54 by Illumina, and 13 by LGTC. Two genes were found significantly differentially expressed by all platforms and the four genes identified by the ABI platform were found by at least three other platforms. Quantitative RT-PCR analysis confirmed 20 out of 28 of the genes detected by two or more platforms and 8 out of 15 of the genes detected by Agilent only. We observed improved correlations between platforms when ranking the genes based on the significance level than with a fixed statistical cut-off. We demonstrate significant overlap in the affected gene sets identified by the different platforms, although biological processes were represented by only partially overlapping sets of genes. Aberrances in GABA-ergic signalling in the transgenic mice were consistently found by all platforms.

Conclusion

The different microarray platforms give partially complementary views on biological processes affected. Our data indicate that when analyzing samples with only subtle differences in gene expression the use of two different platforms might be more attractive than increasing the number of replicates. Commercial two-color platforms seem to have higher power for finding differentially expressed genes between groups with small differences in expression.

A non-parametric meta-analysis approach for combining independent microarray datasets: application using two microarray datasets pertaining to chronic allograft nephropathy

Background

With the popularity of DNA microarray technology, multiple groups of researchers have studied the gene expression of similar biological conditions. Different methods have been developed to integrate the results from various microarray studies, though most of them rely on distributional assumptions, such as the t-statistic based, mixed-effects model, or Bayesian model methods. However, often the sample size for each individual microarray experiment is small. Therefore, in this paper we present a non-parametric meta-analysis approach for combining data from independent microarray studies, and illustrate its application on two independent Affymetrix GeneChip studies that compared the gene expression of biopsies from kidney transplant recipients with chronic allograft nephropathy (CAN) to those with normal functioning allograft.

Results

The simulation study comparing the non-parametric meta-analysis approach to a commonly used t-statistic based approach shows that the non-parametric approach has better sensitivity and specificity. For the application on the two CAN studies, we identified 309 distinct genes that expressed differently in CAN. By applying Fisher's exact test to identify enriched KEGG pathways among those genes called differentially expressed, we found 6 KEGG pathways to be over-represented among the identified genes. We used the expression measurements of the identified genes as predictors to predict the class labels for 6 additional biopsy samples, and the predicted results all conformed to their pathologist diagnosed class labels.

Conclusion

We present a new approach for combining data from multiple independent microarray studies. This approach is non-parametric and does not rely on any distributional assumptions. The rationale behind the approach is logically intuitive and can be easily understood by researchers not having advanced training in statistics. Some of the identified genes and pathways have been reported to be relevant to renal diseases. Further study on the identified genes and pathways may lead to better understanding of CAN at the molecular level.

Interlaboratory evaluation of genomic signatures for predicting carcinogenicity in the rat

The Critical Path Institute recently established the Predictive Safety Testing Consortium, a collaboration between several companies and the U.S. Food and Drug Administration, aimed at evaluating and qualifying biomarkers for a variety of toxicological endpoints. The Carcinogenicity Working Group of the Predictive Safety Testing Consortium has concentrated on sharing data to test the predictivity of two published hepatic gene expression signatures, including the signature by Fielden et al. (2007, Toxicol. Sci. 99, 90-100) for predicting nongenotoxic hepatocarcinogens, and the signature by Nie et al. (2006, Mol. Carcinog. 45, 914-933) for predicting nongenotoxic carcinogens. Although not a rigorous prospective validation exercise, the consortium approach created an opportunity to perform a meta-analysis to evaluate microarray data from short-term rat studies on over 150 compounds. Despite significant differences in study designs and microarray platforms between laboratories, the signatures proved to be relatively robust and more accurate than expected by chance. The accuracy of the Fielden et al. signature was between 63 and 69%, whereas the accuracy of the Nie et al. signature was between 55 and 64%. As expected, the predictivity was reduced relative to internal validation estimates reported under identical test conditions. Although the signatures were not deemed suitable for use in regulatory decision making, they were deemed worthwhile in the early assessment of drugs to aid decision making in drug development. These results have prompted additional efforts to rederive and evaluate a QPCR-based signature using these samples. When combined with a standardized test procedure and prospective interlaboratory validation, the accuracy and potential utility in preclinical applications can be ascertained.

A genome-wide expression analysis in blood identifies pre-elafin as a biomarker in ARDS

Previous microarray-based studies of acute respiratory distress syndrome (ARDS) were performed using various models to mimic disease pathogenesis. The complexity of the pathophysiologic response to direct or indirect lung injury in ARDS is difficult to reconstruct in experimental conditions. Thus, direct analysis of ARDS patient blood may provide valuable information. We investigated genome-wide gene expression profiles in paired whole blood samples from patients with ARDS (n = 8) during the acute stage (within 3 d of diagnosis) and recovery stage of ARDS (around ICU discharge). Among 126 differentially expressed genes, peptidase inhibitor 3 (PI3, encoding elafin, a potent neutrophil elastase inhibitor) had the largest fold-change (-3-fold changes, acute stage/recovery stage) in expression, indicating down-regulation during the acute stage of ARDS. We further examined plasma PI3 levels in 40 patients with ARDS and 23 at-risk control subjects from the same cohort. There was a coincidence of the microarray findings of lower PI3 gene expression with the lower plasma PI3 during the acute-stage. The plasma PI3 levels were statistically significant different among pre-diagnosis, day of diagnosis, and post-diagnosis groups (ANOVA, P = 0.001), with a trend of decreasing from pre- to post-diagnosis group. The time course of plasma PI3 decrease is well correlated with the course of early ARDS development (Pearson correlation coefficient: -0.52, P = 0.0006). Considering that PI3 can covalently binding to extracellular matrix in lung, circulating PI3 may provide a useful clinical marker for monitoring the early development of ARDS and may have implications for ARDS treatment.

Gene expression microarray data analysis demystified

The increasing use of gene expression microarrays, and depositing of the resulting data into public repositories, means that more investigators are interested in using the technology either directly or through meta analysis of the publicly available data. The tools available for data analysis have generally been developed for use by experts in the field, making them difficult to use by the general research community. For those interested in entering the field, especially those without a background in statistics, it is difficult to understand why experimental results can be so variable. The purpose of this review is to go through the workflow of a typical microarray experiment, to show that decisions made at each step, from choice of platform through statistical analysis methods to biological interpretation, are all sources of this variability.

Public consortium efforts in toxicogenomics

Public consortia provide a forum for addressing questions requiring more resources than one organization alone could bring to bear and engaging many sectors of the scientific community. They are particular well suited for tackling some of the questions encountered in the field of toxicogenomics, where the number of studies and microarray analyses would be prohibitively expensive for a single organization to carry out. Five consortia that stand out in the field of toxicogenomics are the Institutional Life Sciences Institute (ILSI) Health and Environmental Sciences Institute (HESI) Committee on the Application of Genomics to Mechanism Based Risk Assessment, the Toxicogenomics Research Consortium, the MicroArray Quality Control (MAQC) Consortium, the InnoMed PredTox effort, and the Predictive Safety Testing Consortium. Collectively, these consortia efforts have addressed issues such as reproducibility of microarray results, standard practice for assays and analysis, relevance of microarray results to conventional end points, and robustness of statistical models on diverse data sets. Their results demonstrate the impact that the pooling of resources, experience, expertise, and insight found in consortia can have.

Toxicogenomics and classic toxicology: how to improve prediction and mechanistic understanding of human toxicity

The field of toxicogenomics has been advancing during the past decade or so since its origin. Most pharmaceutical companies are using it in one or more ways to improve their productivity and supplement their classic toxicology studies. Acceptance of toxicogenomics will continue to grow as regulatory concerns are addressed, proof of concept studies are disseminated more fully, and internal case studies show value for the use of this new technology in concert with classic testing.

Microarray analysis of gene expression of cancer to guide the use of chemotherapeutics

The beauty of microarray analysis of gene expression (MAGE) is that it can be used to discover some genes that were previously thought to be unrelated to a physiologic or pathologic event. During the period from 1999 to 2007, applications of MAGE in cancer investigation have shifted from molecular profiling, identifying previously undiscovered cancer types, predicting outcomes of cancer patients, revealing metastasis signatures of solid tumors, to guiding the use of therapeutics. The roles of cancer genomic signatures have evolved through three phases. In the first phase, genomic signatures were described in stored cancer specimens and dubbed as molecular portraits of cancer. When gene expression profiles were carefully correlated with sufficient clinical information of cancer patients, new subgroups of cancers with distinct outcomes were revealed. In studies of the second phase, validation of cancer signatures was emphasized and commonly performed with independent groups of cancer specimens or independent data set. In the third phase, cancer genomic signatures have been further expanded beyond depicting the molecular portrait of cancer to predicting patient outcomes and guiding the use of cancer therapeutics. Cancer genomic signatures have become an essential part of a new generation of cancer clinical trials. It is advocated that, in future clinical trials of cancer therapy, the cancer specimens of each participant should be tested for currently available predictor genomic signatures, so that the most effective treatment with the least adverse effects for each patient can be identified. Then, participants can be triaged to an appropriate study group.

Application of a correlation correction factor in a microarray cross-platform reproducibility study

BACKGROUND

Recent research examining cross-platform correlation of gene expression intensities has yielded mixed results. In this study, we demonstrate use of a correction factor for estimating cross-platform correlations.

RESULTS

In this paper, three technical replicate microarrays were hybridized to each of three platforms. The three platforms were then analyzed to assess both intra- and cross-platform reproducibility. We present various methods for examining intra-platform reproducibility. We also examine cross-platform reproducibility using Pearson's correlation. Additionally, we previously developed a correction factor for Pearson's correlation which is applicable when X and Y are measured with error. Herein we demonstrate that correcting for measurement error by estimating the "disattenuated" correlation substantially improves cross-platform correlations.

CONCLUSION

When estimating cross-platform correlation, it is essential to thoroughly evaluate intra-platform reproducibility as a first step. In addition, since measurement error is present in microarray gene expression data, methods to correct for attenuation are useful in decreasing the bias in cross-platform correlation estimates.

DeltaFosB induction in orbitofrontal cortex mediates tolerance to cocaine-induced cognitive dysfunction

Current cocaine users show little evidence of cognitive impairment and may perform better when using cocaine, yet withdrawal from prolonged cocaine use unmasks dramatic cognitive deficits. It has been suggested that such impairments arise in part through drug-induced dysfunction within the orbitofrontal cortex (OFC), yet the neurobiological mechanisms remain unknown. We observed that chronic cocaine self-administration increased expression of the transcription factor deltaFosB within both medial and orbitofrontal regions of the rat prefrontal cortex. However, the increase in OFC deltaFosB levels was more pronounced after self-administered rather than experimenter-administered cocaine, a pattern that was not observed in other regions. We then used rodent tests of attention and decision making to determine whether deltaFosB within the OFC contributes to drug-induced alterations in cognition. Chronic cocaine treatment produced tolerance to the cognitive impairments caused by acute cocaine. Overexpression of a dominant-negative antagonist of deltaFosB, deltaJunD, in the OFC prevented this behavioral adaptation, whereas locally overexpressing deltaFosB mimicked the effects of chronic cocaine. Gene microarray analyses identified potential molecular mechanisms underlying this behavioral change, including an increase in transcription of metabotropic glutamate receptor subunit 5 and GABA(A) receptors as well as substance P. Identification of deltaFosB in the OFC as a mediator of tolerance to the effects of cocaine on cognition provides fundamentally new insight into the transcriptional modifications associated with addiction.

Isolation and molecular characterization of cancer stem cells in MMTV-Wnt-1 murine breast tumors

In human breast cancers, a phenotypically distinct minority population of tumorigenic (TG) cancer cells (sometimes referred to as cancer stem cells) drives tumor growth when transplanted into immunodeficient mice. Our objective was to identify a mouse model of breast cancer stem cells that could have relevance to the study of human breast cancer. To do so, we used breast tumors of the mouse mammary tumor virus (MMTV)-Wnt-1 mice. MMTV-Wnt-1 breast tumors were harvested, dissociated into single-cell suspensions, and sorted by flow cytometry on Thy1, CD24, and CD45. Sorted cells were then injected into recipient background FVB/NJ female syngeneic mice. In six of seven tumors examined, Thy1+CD24+ cancer cells, which constituted approximately 1%-4% of tumor cells, were highly enriched for cells capable of regenerating new tumors compared with cells of the tumor that did not fit this profile ("not-Thy1+CD24+"). Resultant tumors had a phenotypic diversity similar to that of the original tumor and behaved in a similar manner when passaged. Microarray analysis comparing Thy1+CD24+ tumor cells to not-Thy1+CD24+ cells identified a list of differentially expressed genes. Orthologs of these differentially expressed genes predicted survival of human breast cancer patients from two different study groups. These studies suggest that there is a cancer stem cell compartment in the MMTV-Wnt-1 murine breast tumor and that there is a clinical utility of this model for the study of cancer stem cells.

Gene expression changes induced by the tumorigenic pyrrolizidine alkaloid riddelliine in liver of Big Blue rats

Background

Pyrrolizidine alkaloids (PAs) are probably the most common plant constituents that poison livestock, wildlife, and humans worldwide. Riddelliine is isolated from plants grown in the western United States and is a prototype of genotoxic PAs. Riddelliine was used to investigate the genotoxic effects of PAs via analysis of gene expression in the target tissue of rats in this study. Previously we observed that the mutant frequency in the liver of rats gavaged with riddelliine was 3-fold higher than that in the control group. Molecular analysis of the mutants indicated that there was a statistically significant difference between the mutational spectra from riddelliine-treated and control rats.

Results

Riddelliine-induced gene expression profiles in livers of Big Blue transgenic rats were determined. The female rats were gavaged with riddelliine at a dose of 1 mg/kg body weight 5 days a week for 12 weeks. Rat whole genome microarray was used to perform genome-wide gene expression studies. When a cutoff value of a two-fold change and a P-value less than 0.01 were used as gene selection criteria, 919 genes were identified as differentially expressed in riddelliine-treated rats compared to the control animals. By analysis with the Ingenuity Pathway Analysis Network, we found that these significantly changed genes were mainly involved in cancer, cell death, tissue development, cellular movement, tissue morphology, cell-to-cell signaling and interaction, and cellular growth and proliferation. We further analyzed the genes involved in metabolism, injury of endothelial cells, liver abnormalities, and cancer development in detail.

Conclusion

The alterations in gene expression were directly related to the pathological outcomes reported previously. These results provided further insight into the mechanisms involved in toxicity and carcinogenesis after exposure to riddelliine, and permitted us to investigate the interaction of gene products inside the signaling networks.

Reproducibility of microarray data: a further analysis of microarray quality control (MAQC) data

BACKGROUND

Many researchers are concerned with the comparability and reliability of microarray gene expression data. Recent completion of the MicroArray Quality Control (MAQC) project provides a unique opportunity to assess reproducibility across multiple sites and the comparability across multiple platforms. The MAQC analysis presented for the conclusion of inter- and intra-platform comparability/reproducibility of microarray gene expression measurements is inadequate. We evaluate the reproducibility/comparability of the MAQC data for 12901 common genes in four titration samples generated from five high-density one-color microarray platforms and the TaqMan technology. We discuss some of the problems with the use of correlation coefficient as metric to evaluate the inter- and intra-platform reproducibility and the percent of overlapping genes (POG) as a measure for evaluation of a gene selection procedure by MAQC.

RESULTS

A total of 293 arrays were used in the intra- and inter-platform analysis. A hierarchical cluster analysis shows distinct differences in the measured intensities among the five platforms. A number of genes show a small fold-change in one platform and a large fold-change in another platform, even though the correlations between platforms are high. An analysis of variance shows thirty percent of gene expressions of the samples show inconsistent patterns across the five platforms. We illustrated that POG does not reflect the accuracy of a selected gene list. A non-overlapping gene can be truly differentially expressed with a stringent cut, and an overlapping gene can be non-differentially expressed with non-stringent cutoff. In addition, POG is an unusable selection criterion. POG can increase or decrease irregularly as cutoff changes; there is no criterion to determine a cutoff so that POG is optimized.

CONCLUSION

Using various statistical methods we demonstrate that there are differences in the intensities measured by different platforms and different sites within platform. Within each platform, the patterns of expression are generally consistent, but there is site-by-site variability. Evaluation of data analysis methods for use in regulatory decision should take no treatment effect into consideration, when there is no treatment effect, "a fold-change cutoff with a non-stringent p-value cutoff" could result in 100% false positive error selection.

Extended analysis of benchmark datasets for Agilent two-color microarrays

Background

As part of its broad and ambitious mission, the MicroArray Quality Control (MAQC) project reported the results of experiments using External RNA Controls (ERCs) on five microarray platforms. For most platforms, several different methods of data processing were considered. However, there was no similar consideration of different methods for processing the data from the Agilent two-color platform. While this omission is understandable given the scale of the project, it can create the false impression that there is consensus about the best way to process Agilent two-color data. It is also important to consider whether ERCs are representative of all the probes on a microarray.

Results

A comparison of different methods of processing Agilent two-color data shows substantial differences among methods for low-intensity genes. The sensitivity and specificity for detecting differentially expressed genes varies substantially for different methods. Analysis also reveals that the ERCs in the MAQC data only span the upper half of the intensity range, and therefore cannot be representative of all genes on the microarray.

Conclusion

Although ERCs demonstrate good agreement between observed and expected log-ratios on the Agilent two-color platform, such an analysis is incomplete. Simple loess normalization outperformed data processing with Agilent's Feature Extraction software for accurate identification of differentially expressed genes. Results from studies using ERCs should not be over-generalized when ERCs are not representative of all probes on a microarray.

Macromolecule biosynthesis: a key function of sleep

The function(s) of sleep remains a major unanswered question in biology. We assessed changes in gene expression in the mouse cerebral cortex and hypothalamus following different durations of sleep and periods of sleep deprivation. There were significant differences in gene expression between behavioral states; we identified 3,988 genes in the cerebral cortex and 823 genes in the hypothalamus with altered expression patterns between sleep and sleep deprivation. Changes in the steady-state level of transcripts for various genes are remarkably common during sleep, as 2,090 genes in the cerebral cortex and 409 genes in the hypothalamus were defined as sleep specific and changed (increased or decreased) their expression during sleep. The largest categories of overrepresented genes increasing expression with sleep were those involved in biosynthesis and transport. In both the cerebral cortex and hypothalamus, during sleep there was upregulation of multiple genes encoding various enzymes involved in cholesterol synthesis, as well as proteins for lipid transport. There was also upregulation during sleep of genes involved in synthesis of proteins, heme, and maintenance of vesicle pools, as well as antioxidant enzymes and genes encoding proteins of energy-regulating pathways. We postulate that during sleep there is a rebuilding of multiple key cellular components in preparation for subsequent wakefulness.

Identification of activity-dependent gene expression profiles reveals specific subsets of genes induced by different routes of Ca(2+) entry in cultured rat cortical neurons

Neuronal activity-dependent gene transcription is a key feature of long-lasting synaptic strengthening associated with learning and memory, as well as activity-dependent neuroprotection. To comprehensively determine the molecular alterations, we carried out genome-wide microarray analysis in cultured rat cortical neurons treated with specific pharmacological agents, a model with alterations in neuronal activity, which were monitored by multi-site electrophysiological recordings. Of the approximately 27,000 genes, the expression of 248 genes was strongly changed in response to enhanced activity. These genes encompass a large number of members of distinct families, including synaptic vesicle proteins, ion channels, signal transduction molecules, synaptic growth regulators, and others. Two subsets of these genes were further confirmed to be specifically induced by Ca(2+) influx through N-methyl-D-aspartate (NMDA) receptors and L-type voltage-gated Ca(2+) channels (VGCCs). In addition, those genes dynamically regulated by the enhanced activity were also elucidated, as well as those candidate genes associated with synaptic plasticity and neuroprotection. Our findings therefore would help define the molecular mechanisms that occur in response to neuronal activity and identify specific clusters of genes that contribute to activity-dependent and Ca(2+)-inducible modulation of brain development and function.

Taking gene-expression profiling to the clinic: when will molecular signatures become relevant to patient care?

The advent of microarray technology has enabled scientists to simultaneously investigate the expression of thousands of genes. Gene-expression profiling studies have provided a molecular classification of breast cancer into clinically relevant subtypes, new tools to predict disease recurrence and response to different treatments, and new insights into various oncogenic pathways and the process of metastatic progression. Here we describe the state of the art of gene-expression studies in breast cancer, and consider both their current limitations and future promises. We also discuss the potential of molecular signatures to have an impact on individual breast cancer patient management, and ultimately to accelerate the transition between empirical and tailored medicine.

New data on robustness of gene expression signatures in leukemia: comparison of three distinct total RNA preparation procedures

BACKGROUND

Microarray gene expression (MAGE) signatures allow insights into the transcriptional processes of leukemias and may evolve as a molecular diagnostic test. Introduction of MAGE into clinical practice of leukemia diagnosis will require comprehensive assessment of variation due to the methodologies. Here we systematically assessed the impact of three different total RNA isolation procedures on variation in expression data: method A: lysis of mononuclear cells, followed by lysate homogenization and RNA extraction; method B: organic solvent based RNA isolation, and method C: organic solvent based RNA isolation followed by purification.

RESULTS

We analyzed 27 pediatric acute leukemias representing nine distinct subtypes and show that method A yields better RNA quality, was associated with more differentially expressed genes between leukemia subtypes, demonstrated the lowest degree of variation between experiments, was more reproducible, and was characterized with a higher precision in technical replicates. Unsupervised and supervised analyses grouped leukemias according to lineage and clinical features in all three methods, thus underlining the robustness of MAGE to identify leukemia specific signatures.

CONCLUSION

The signatures in the different subtypes of leukemias, regardless of the different extraction methods used, account for the biggest source of variation in the data. Lysis of mononuclear cells, followed by lysate homogenization and RNA extraction represents the optimum method for robust gene expression data and is thus recommended for obtaining robust classification results in microarray studies in acute leukemias.

Establishing a major cause of discrepancy in the calibration of Affymetrix GeneChips

Background

Affymetrix GeneChips are a popular platform for performing whole-genome experiments on the transcriptome. There are a range of different calibration steps, and users are presented with choices of different background subtractions, normalisations and expression measures. We wished to establish which of the calibration steps resulted in the biggest uncertainty in the sets of genes reported to be differentially expressed.

Results

Our results indicate that the sets of genes identified as being most significantly differentially expressed, as estimated by the z-score of fold change, is relatively insensitive to the choice of background subtraction and normalisation. However, the contents of the gene list are most sensitive to the choice of expression measure. This is irrespective of whether the experiment uses a rat, mouse or human chip and whether the chip definition is made using probe mappings from Unigene, RefSeq, Entrez Gene or the original Affymetrix definitions. It is also irrespective of whether both Present and Absent, or just Present, Calls from the MAS5 algorithm are used to filter genelists, and this conclusion holds for genes of differing intensities. We also reach the same conclusion after assigning genes to be differentially expressed using t-statistics, although this approach results in a large amount of false positives in the sets of genes identified due to the small numbers of replicates typically used in microarray experiments.

Conclusion

The major calibration uncertainty that biologists need to consider when analysing Affymetrix data is how their multiple probe values are condensed into one expression measure.

Robust interlaboratory reproducibility of a gene expression signature measurement consistent with the needs of a new generation of diagnostic tools

Background

The increasing use of DNA microarrays in biomedical research, toxicogenomics, pharmaceutical development, and diagnostics has focused attention on the reproducibility and reliability of microarray measurements. While the reproducibility of microarray gene expression measurements has been the subject of several recent reports, there is still a need for systematic investigation into what factors most contribute to variability of measured expression levels observed among different laboratories and different experimenters.

Results

We report the results of an interlaboratory comparison of gene expression array measurements on the same microarray platform, in which the RNA amplification and labeling, hybridization and wash, and slide scanning were each individually varied. Identical input RNA was used for all experiments. While some sources of variation have measurable influence on individual microarray signals, they showed very low influence on sample-to-reference ratios based on averaged triplicate measurements in the two-color experiments. RNA labeling was the largest contributor to interlaboratory variation.

Conclusion

Despite this variation, measurement of one particular breast cancer gene expression signature in three different laboratories was found to be highly robust, showing a high intralaboratory and interlaboratory reproducibility when using strictly controlled standard operating procedures.

Functional network analysis of the transcriptomes of mesenchymal stem cells derived from amniotic fluid, amniotic membrane, cord blood, and bone marrow

Using high-density oligonucleotide microarrays and functional network analyses, we examined whether MSCs derived from four different origins exhibited unique gene expression profiles individually and then compared the gene expression profiles of all MSCs with those of fetal organs. Our results indicated that within each group of MSCs from the same origin, the variability of the gene expression levels was smaller than that between groups of different origins. Functional genomic studies revealed the specific roles of MSCs from different origins. Our results suggest that amniotic fluid MSCs may initiate interactions with the uterus by upregulating oxytocin and thrombin receptors. Amniotic membrane MSCs may play a role in maintaining homeostasis of fluid and electrolytes by regulating the networks of endothelin, neprilysin, bradykinin receptors, and atrial natriuretic peptide. Cord blood MSCs may be involved in innate immune systems as the neonatal defense system against the earliest encountered pathogens. Adult bone marrow MSCs may be an important source not only of all blood lineages but also of bone formation. However, in spite of the different gene expression profiles seen in MSCs derived from different origins, a set of core gene expression profiles was preserved in these four kinds of MSCs. The core signature transcriptomes of all MSCs, when contrasted against those of fetal organs, included genes involved in the regulation of extracellular matrix and adhesion, transforming growth factor-beta receptor signaling, and the Wnt signaling pathways. Disclosure of potential conflicts of interest is found at the end of this article.

QA/QC issues to aid regulatory acceptance of microarray gene expression data

The U.S. Food and Drug Administration is responsible for (1) promoting and protecting public health by assuring the safety and effectiveness of medicines and medical devices and (2) advancing public health by helping to speed innovations that make medicines and foods safer, more effective, and more affordable. The genomics revolution has dramatically increased our knowledge of basic biology but this has not resulted in the expected acceleration of new medical product development. The Agency's Critical Path to New Medical Products stresses that new tools are needed to address this pipeline problem. Microarray technology is one of these promising tools although questions have risen about the reproducibility of measurements. The Microarray Quality Control (MAQC) Project was initiated by FDA scientists to address this issue. This large project, which evaluated reference RNA samples on seven microarray platforms, found good intralaboratory repeatability and interlaboratory reproducibility. In addition, there was high cross-platform consistency. All data are available free of cost and the reference RNA samples are available for proficiency testing. Thus, current microarray technology appears to provide both reliability and consistency for regulatory submissions.

Review of the literature examining the correlation among DNA microarray technologies

DNA microarray technologies are used in a variety of biological disciplines. The diversity of platforms and analytical methods employed has raised concerns over the reliability, reproducibility and correlation of data produced across the different approaches. Initial investigations (years 2000-2003) found discrepancies in the gene expression measures produced by different microarray technologies. Increasing knowledge and control of the factors that result in poor correlation among the technologies has led to much higher levels of correlation among more recent publications (years 2004 to present). Here, we review the studies examining the correlation among microarray technologies. We find that with improvements in the technology (optimization and standardization of methods, including data analysis) and annotation, analysis across platforms yields highly correlated and reproducible results. We suggest several key factors that should be controlled in comparing across technologies, and are good microarray practice in general.

Resistance exercise reverses aging in human skeletal muscle

Human aging is associated with skeletal muscle atrophy and functional impairment (sarcopenia). Multiple lines of evidence suggest that mitochondrial dysfunction is a major contributor to sarcopenia. We evaluated whether healthy aging was associated with a transcriptional profile reflecting mitochondrial impairment and whether resistance exercise could reverse this signature to that approximating a younger physiological age. Skeletal muscle biopsies from healthy older (N = 25) and younger (N = 26) adult men and women were compared using gene expression profiling, and a subset of these were related to measurements of muscle strength. 14 of the older adults had muscle samples taken before and after a six-month resistance exercise-training program. Before exercise training, older adults were 59% weaker than younger, but after six months of training in older adults, strength improved significantly (P<0.001) such that they were only 38% lower than young adults. As a consequence of age, we found 596 genes differentially expressed using a false discovery rate cut-off of 5%. Prior to the exercise training, the transcriptome profile showed a dramatic enrichment of genes associated with mitochondrial function with age. However, following exercise training the transcriptional signature of aging was markedly reversed back to that of younger levels for most genes that were affected by both age and exercise. We conclude that healthy older adults show evidence of mitochondrial impairment and muscle weakness, but that this can be partially reversed at the phenotypic level, and substantially reversed at the transcriptome level, following six months of resistance exercise training.

Three methods for optimization of cross-laboratory and cross-platform microarray expression data

Microarray gene expression data becomes more valuable as our confidence in the results grows. Guaranteeing data quality becomes increasingly important as microarrays are being used to diagnose and treat patients (1–4). The MAQC Quality Control Consortium, the FDA's Critical Path Initiative, NCI's caBIG and others are implementing procedures that will broadly enhance data quality. As GEO continues to grow, its usefulness is constrained by the level of correlation across experiments and general applicability. Although RNA preparation and array platform play important roles in data accuracy, pre-processing is a user-selected factor that has an enormous effect. Normalization of expression data is necessary, but the methods have specific and pronounced effects on precision, accuracy and historical correlation. As a case study, we present a microarray calibration process using normalization as the adjustable parameter. We examine the impact of eight normalizations across both Agilent and Affymetrix expression platforms on three expression readouts: (1) sensitivity and power, (2) functional/biological interpretation and (3) feature selection and classification error. The reader is encouraged to measure their own discordant data, whether cross-laboratory, cross-platform or across any other variance source, and to use their results to tune the adjustable parameters of their laboratory to ensure increased correlation.