Analysing gene expression data from DNA microarrays to identify candidate genes

An investigation into the detoxification of microcystin-LR by the glutathione pathway in Balb/c mice

Toxin-producing cyanobacteria pose a world-wide health threat to humans and animals due to their increasing presence in both drinking and recreational waters. The predominant cyanotoxin, microcystin-LR (MCLR), targets the liver and its toxicity depends on the uptake and removal rates in the liver. The role of the glutathione detoxification pathway in protecting the liver from the effects of MCLR was investigated. Mice exposed to a single 75% LD(50) dose of pure MCLR were sacrificed at 8, 16, 24 and 32 h post-exposure (pe). Toxin induced liver damage was observed 8 and 16 h pe as evidenced by raised serum ALT and LDH levels, reduced glycogen levels and liver histology. A significant increase in lipid peroxidation was seen at 16 h pe that decreased after 24 and 32 h pe, the time-points which showed significant increases in GPX activity. An increase in soluble GST activity was noted between 8 and 16 h pe, levels of total GSH increased at 24 h while oxidised glutathione increased throughout the investigation. The increase in activity of both GPX and GST corresponded with increased transcription of these enzymes, as well as the rate-limiting enzyme in GSH synthesis, gamma-glutamyl transferase. In conclusion, this study confirms that an increase in GST activity is critical for the detoxification of MCLR, that this is regulated at the transcriptional level, and that exposure to MCLR induces the de novo synthesis of GSH. Finally, we report the involvement of GPX in the removal of MCLR-induced lipid hydroperoxides.

Conserved mechanisms across development and tumorigenesis revealed by a mouse development perspective of human cancers

Identification of common mechanisms underlying organ development and primary tumor formation should yield new insights into tumor biology and facilitate the generation of relevant cancer models. We have developed a novel method to project the gene expression profiles of medulloblastomas (MBs)--human cerebellar tumors--onto a mouse cerebellar development sequence: postnatal days 1-60 (P1-P60). Genomically, human medulloblastomas were closest to mouse P1-P10 cerebella, and normal human cerebella were closest to mouse P30-P60 cerebella. Furthermore, metastatic MBs were highly associated with mouse P5 cerebella, suggesting that a clinically distinct subset of tumors is identifiable by molecular similarity to a precise developmental stage. Genewise, down- and up-regulated MB genes segregate to late and early stages of development, respectively. Comparable results for human lung cancer vis-a-vis the developing mouse lung suggest the generalizability of this multiscalar developmental perspective on tumor biology. Our findings indicate both a recapitulation of tissue-specific developmental programs in diverse solid tumors and the utility of tumor characterization on the developmental time axis for identifying novel aspects of clinical and biological behavior.

Comparing genomic and histologic correlations to radiographic changes in tumors: a murine SCC VII model study

RATIONALE AND OBJECTIVES

To investigate the correlation between the temporal changes in T1- and T2-weighted contrast-enhanced magnetic resonance imaging (MRI), histologic evaluation, and genomic analysis using oligonucleotide microarrays in a murine squamous cell carcinoma tumor models.

MATERIALS AND METHODS

The squamous cell carcinoma (SCC VII) cell line was used to initiate subcutaneous tumors in mice. This mouse model has been used as a model for human head and neck carcinomas. Animals were imaged using contrast enhanced MRI (CE-MRI). Different stages of tumor growth were defined based on changes in the T1- and T2-weighted MRI patterns. The contrast enhancing (CE) and nonenhancing (NE) regions of the tumors were marked and biopsied for oligonucleotide microarray and histologic analysis. Tumors with no differential contrast enhancement were used as controls.

RESULTS

Distinct temporal stages of tumor progression can be defined using both T1- and T2-weighted CE-MRI and microarray analysis. The early stage tumors show a homogeneous contrast enhancement pattern in the T1- and T2-weighted images with no significant differential gene expression from the center and periphery of the tumor. The more advanced tumors that show discrete regions of contrast enhancement in the post-contrast T1-weighted MRIs and tissues from the CE and NE regions show distinctly differential gene expression profiles. Histologic analysis (hematoxylin-eosin stain) showed that the samples obtained from the periphery and center of the early stage tumors and the CE and NE regions from these more advanced tumors were similar. The gene expression profiles of late-stage tumors that showed changes in T2-weighted MRI signal intensity were consistent with tissue degradation in the NE region, which also showed characteristic signs of tissue necrosis in histologic analysis.

CONCLUSION

These results show that temporal changes in T1- and T2-weighted CE-MRI are related to distinct gene expression profiles, and histologic analysis may not be sufficient to detect these detailed changes. As tumors progress, discrete regions of post-contrast T1 enhancement are identified; these regions have distinct gene expression patterns despite similar histologic features. In late-stage tumors, regions of T2 signal changes are observed which correspond with tissue necrosis.

Emerging genetic technologies in clinical and research settings

With the rapid expansion of genomic health care, nurses are exposed to emerging genetic technologies in a wide variety of clinical and research settings; however, nurses have limited knowledge about these technologies. The polymerase chain reaction procedure, which is the foundation of current molecular genetic technologies, real-time polymerase chain reaction, and microarray analysis are described in this article. The applications, strengths, and limitations of each technology are discussed.

Sensitivity and reproducibility of standardized-competitive RT-PCR for transcript quantification and its comparison with real time RT-PCR

Background

Probe based detection assays form the mainstay of transcript quantification. Problems with these assays include varying hybridization efficiencies of the probes used for transcript quantification and the expense involved. We examined the ability of a standardized competitive RT-PCR (StaRT PCR) assay to quantify transcripts of 4 cell cycle associated genes (RB, E2F1, CDKN2A and PCNA) in two cell lines (T24 & LD419) and compared its efficacy with the established Taqman real time quantitative RT-PCR assay. We also assessed the sensitivity, reproducibility and consistency of StaRT PCR. StaRT PCR assay is based on the incorporation of competitive templates (CT) in precisely standardized quantities along with the native template (NT) in a PCR reaction. This enables transcript quantification by comparing the NT and CT band intensities at the end of the PCR amplification. The CT serves as an ideal internal control. The transcript numbers are expressed as copies per million transcripts of a control gene such as β-actin (ACTB).

Results

The NT and CT were amplified at remarkably similar rates throughout the StaRT PCR amplification cycles, and the coefficient of variation was least (<3.8%) when the NT/CT ratio was kept as close to 1:1 as possible. The variability between the rates of amplification in different tubes subjected to the same StaRT PCR reaction was very low and within the range of experimental noise. Further, StaRT PCR was sensitive enough to detect variations as low as 10% in endogenous actin transcript quantity (p < 0.01 by the paired student's t-test). StaRT PCR correlated well with Taqman real time RT-PCR assay in terms of transcript quantification efficacy (p < 0.01 for all 4 genes by the Spearman Rank correlation method) and the ability to discriminate between cell types and confluence patterns.

Conclusion

StaRT PCR is thus a reliable and sensitive technique that can be applied to medium-high throughput quantitative transcript measurement. Further, it correlates well with Taqman real time PCR in terms of quantitative and discriminatory ability. This label-free, inexpensive technique may provide the ability to generate prognostically important molecular signatures unique to individual tumors and may enable identification of novel therapeutic targets.

Identification of novel proteins associated with hepatocellular carcinomas using protein microarrays

Characterization of the protein profiles expressed by hepatocellular carcinomas (HCCs) may identify the genes involved in hepatocellular carcinogenesis and offers the possibility of elucidating clinical biomarkers. In an effort to discover such proteins and pathways that are deregulated in hepatocellular carcinogenesis, cellular proteomes of matched normal liver cells and carcinoma were analysed by tissue microdissection and protein microarrays. Using protein microarrays made up of 83 different antibodies, it was possible to monitor alterations of the protein levels in HCC and non-neoplastic liver tissue. Further analysis of altered proteins was performed using western blot analysis and tissue microarrays (TMAs) containing 210 HCC specimens and corresponding liver tissue. The protein microarray approach revealed differential expression between HCC and normal liver of 32 of the 83 proteins examined: 21 of these were up-regulated and 11 down-regulated. IGF (insulin growth factor) II, ADAM (a disintegrin and metalloproteases) 9, STAT (signal transducers and activators of transcription) 3, SOCS (suppressors of cytokine signalling) 3, and cyclin D1 were significantly up-regulated and collagen I, SMAD 4, FHIT (fragile histidine triad), and SOCS1 were down-regulated. The differential expression of these proteins was confirmed using western blot analysis and TMAs. Correlation of differentially regulated proteins with clinico-pathological data showed that cyclin D1 and SOCS1 were associated with tumour prognosis in univariate analysis, but not multivariate analysis. These data indicate that the development of an array-based approach for the determination of protein profiles in HCC may facilitate the identification of new proteins associated with carcinogenesis or prognosis.

Fundamentals of cDNA microarray data analysis

Microarray technology is a powerful approach for genomics research. The multi-step, data-intensive nature of this technology has created an unprecedented informatics and analytical challenge. It is important to understand the crucial steps that can affect the outcome of the analysis. In this review, we provide an overview of the contemporary trend on various main analysis steps in the microarray data analysis process, which includes experimental design, data standardization, image acquisition and analysis, normalization, statistical significance inference, exploratory data analysis, class prediction and pathway analysis, as well as various considerations relevant to their implementation.

Pancreatic cancer escape variants that evade immunogene therapy through loss of sensitivity to IFNgamma-induced apoptosis

Combined injections into experimental tumor nodules of adenovirus encoding IL-12 and certain chemokines are capable to induce immune-mediated complete regressions. In this study, we found that the combination of two adenoviruses, one encoding IL-12 and other MIP3alpha (AdCMVIL-12+AdCMVMIP3alpha) was very successful in treating CT-26-derived colon carcinomas. However, in experimental tumors generated from the pancreatic carcinoma cell line Panc02 such combined treatment induces 50% of macroscopic complete regressions, although local relapses within 1 week are almost constant. We derived cell lines from such relapsing tumors and found that experimental malignancies derived from their inoculum were not amenable to treatment in any case with AdCMVIL-12+AdCMVMIP-3alpha. Importantly, relapsing cell lines were insensitive to in vitro induction of apoptosis by IFNgamma, in clear contrast with the original Panc02 cells. Comparative analyses by cDNA arrays of relapsing cell lines versus wild-type Panc02 were performed revealing an important number of genes (383) whose expression levels were modified more than two-fold. These changes grouped in certain gene ontology categories should harbor the mechanistic explanations of the acquired selective resistance to IFNgamma.

Identification of genes differentially expressed during aflatoxin biosynthesis in Aspergillus flavus and Aspergillus parasiticus

A complex regulatory network governs the biosynthesis of aflatoxin. While several genes involved in aflatoxin production are known, their action alone cannot account for its regulation. Arrays of clones from an Aspergillus flavus cDNA library and glass slide microarrays of ESTs were screened to identify additional genes. An initial screen of the cDNA clone arrays lead to the identification of 753 unique ESTs. Many showed sequence similarity to known metabolic and regulatory genes; however, no function could be ascribed to over 50% of the ESTs. Gene expression analysis of Aspergillus parasiticus grown under conditions conducive and non-conductive for aflatoxin production was evaluated using glass slide microarrays containing the 753 ESTs. Twenty-four genes were more highly expressed during aflatoxin biosynthesis and 18 genes were more highly expressed prior to aflatoxin biosynthesis. No predicted function could be ascribed to 18 of the 24 genes whose elevated expression was associated with aflatoxin biosynthesis.

The design and analysis of microarray experiments: applications in parasitology

Microarray experiments can generate enormous amounts of data, but large datasets are usually inherently complex, and the relevant information they contain can be difficult to extract. For the practicing biologist, we provide an overview of what we believe to be the most important issues that need to be addressed when dealing with microarray data. In a microarray experiment we are simply trying to identify which genes are the most "interesting" in terms of our experimental question, and these will usually be those that are either overexpressed or underexpressed (upregulated or downregulated) under the experimental conditions. Analysis of the data to find these genes involves first preprocessing of the raw data for quality control, including filtering of the data (e.g., detection of outlying values) followed by standardization of the data (i.e., making the data uniformly comparable throughout the dataset). This is followed by the formal quantitative analysis of the data, which will involve either statistical hypothesis testing or multivariate pattern recognition. Statistical hypothesis testing is the usual approach to "class comparison," where several experimental groups are being directly compared. The best approach to this problem is to use analysis of variance, although issues related to multiple hypothesis testing and probability estimation still need to be evaluated. Pattern recognition can involve "class prediction," for which a range of supervised multivariate techniques are available, or "class discovery," for which an even broader range of unsupervised multivariate techniques have been developed. Each technique has its own limitations, which need to be kept in mind when making a choice from among them. To put these ideas in context, we provide a detailed examination of two specific examples of the analysis of microarray data, both from parasitology, covering many of the most important points raised.

Microarrays: new tools for transplantation research

The advent of DNA microarray technology has greatly enhanced our potential to understand the molecular basis of human diseases and to aid in more accurate classification, diagnosis and/or prognosis. This powerful, flexible and highly informative technique has been adopted by many biomedical research disciplines. The use of DNA microarrays for gene expression profiling of patients undergoing organ transplantation has diagnostic and therapeutic potential. By generating global views of the gene expression changes in renal graft function post transplantation, DNA microarray technology will provide important information to improve our understanding of the molecular basis of various causes of graft dysfunction, and therefore suggest improved diagnosis, disease classification, and treatment.

Temporal gene expression profile of human precursor B leukemia cells induced by adhesion receptor: identification of pathways regulating B-cell survival

The physical interactions between B cells and stromal cells from the lymphoid tissue microenvironment are critical to the survival of normal and malignant B cells. They are principally mediated by integrins expressed on B cells and counterreceptors on stromal cells. Specifically, alpha4beta1 integrin engagement rescues B cells from physiological or drug-induced apoptosis. Therefore, in order to understand the mechanisms by which integrins prevent apoptosis in leukemia B cells, we compared the temporal gene expression profiles induced by beta1-integrin ligation with fibronectin (Fn) or adhesion by poly-L-Lysine in serum-starved precursor B leukemia cells. Among the 38 selected differentially expressed genes, 6 genes involved in adhesion (VAV2, EPB41L1, CORO1A), proliferation (FRAP1, CCT4), and intercellular communication (GJB3) were validated by real-time quantitative polymerase chain reaction (RT-Q-PCR). Gene expression modulation could also be validated at the protein level for 5 other genes. We show that integrin stimulation up-regulated FBI-1 expression but inhibited CD79a, Requiem, c-Fos, and caspase 7 induction when the cells underwent apoptosis. We further demonstrate that Fn stimulation also inhibits caspase 3 activation but increases XIAP and survivin expression. Moreover, integrin stimulation also prevents caspase activation induced by doxorubicin. Therefore, we identified genes modulated by adhesion of human precursor B leukemia cells that regulate proliferation and apoptosis, highlighting new pathways that might provide insights into future therapy aiming at targeting apoptosis of leukemia cells.

Receiver operating characteristic analysis: a general tool for DNA array data filtration and performance estimation

A critical step for DNA array analysis is data filtration, which can reduce thousands of detected signals to limited sets of genes. Commonly accepted rules for such filtration are still absent. We present a rational approach, based on thresholding of intensities with cutoff levels that are estimated by receiver operating characteristic (ROC) analysis. The technique compares test results with known distributions of positive and negative signals. We apply the method to Atlas cDNA arrays, GeneFilters, and Affymetrix GeneChip. ROC analysis demonstrates similarities in the distribution of false and true positive data for these different systems. We illustrate the estimation of an optimal cutoff level for intensity-based filtration, providing the highest ratio of true to false signals. For GeneChip arrays, we derived filtration thresholds consistent with the reported data based on replicate hybridizations. Intensity-based filtration optimized with ROC combined with other types of filtration (for example, based on significances of differences and/or ratios), should improve DNA array analysis. ROC methodology is also demonstrated for comparison of the performance of different types of arrays, imagers, and analysis software.

Primer on medical genomics part V: bioinformatics

Bioinformatics is the discipline that develops and applies informatics to the field of molecular biology. Although a comprehensive review of the entire field of bioinformatics is beyond the scope of this article, I review the basic tenets of the field and provide a topical sampling of the popular technologies available to clinicians and researchers. These technologies include tools and methods for sequence analysis (nucleotide and protein sequences), rendering of secondary and tertiary structures for these molecules, and protein fold prediction that can lead to rational drug design. I then discuss signaling pathways, new standards for data representation of genes and proteins, and finally the promise of merging these molecular data with the clinical world (the new science of phenomics).

The use and analysis of microarray data

Functional genomics is the study of gene function through the parallel expression measurements of genomes, most commonly using the technologies of microarrays and serial analysis of gene expression. Microarray usage in drug discovery is expanding, and its applications include basic research and target discovery, biomarker determination, pharmacology, toxicogenomics, target selectivity, development of prognostic tests and disease-subclass determination. This article reviews the different ways to analyse large sets of microarray data, including the questions that can be asked and the challenges in interpreting the measurements.

Reproducibility of oligonucleotide microarray transcriptome analyses. An interlaboratory comparison using chemostat cultures of Saccharomyces cerevisiae

Assessment of reproducibility of DNA-microarray analysis from published data sets is complicated by the use of different microbial strains, cultivation techniques, and analytical procedures. Because intra- and interlaboratory reproducibility is highly relevant for application of DNA-microarray analysis in functional genomics and metabolic engineering, we designed a set of experiments to specifically address this issue. Saccharomyces cerevisiae CEN.PK113-7D was grown under defined conditions in glucose-limited chemostats, followed by transcriptome analysis with Affymetrix GeneChip arrays. In each of the laboratories, three independent replicate cultures were grown aerobically as well as anaerobically. Although variations introduced by in vitro handling steps were small and unbiased, greater variation from replicate cultures underscored that, to obtain reliable information, experimental replication is essential. Under aerobic conditions, 86% of the most highly expressed yeast genes showed an average intralaboratory coefficient of variation of 0.23. This is significantly lower than previously reported for shake-flask-culture transcriptome analyses and probably reflects the strict control of growth conditions in chemostats. Using the triplicate data sets and appropriate statistical analysis, the change calls from anaerobic versus aerobic comparisons yielded an over 95% agreement between the laboratories for transcripts that changed by over 2-fold, leaving only a small fraction of genes that exhibited laboratory bias.

Gene-expression profiling of experimental autoimmune encephalomyelitis

Experimental autoimmune encephalomyelitis (EAE) is a mouse model that serves as an experimental tool for studying the etiology, pathogenesis, as well as new therapeutic approaches of multiple sclerosis (MS). EAE is a polygenic chronic inflammatory demyelinating disease of the nervous system that involves the interaction between genetic and environmental factors. Previous studies have identified multiple quantitative trait loci (QTL) controlling different aspects of disease pathogenesis. However, progress in identifying new susceptibility genes outside the MHC locus has been slow. With the advent of new global methods for genetic analysis such as large-scale sequencing, gene expression profiling combined with classic linkage analysis and congenic and physical mapping progress is considerably accelerating. Here we review our preliminary work on the use of gene expression mapping to identify new putative genetic pathways contributing to the pathogenesis of EAE.

Assessment of differential gene expression in human peripheral nerve injury

BACKGROUND

Microarray technology is a powerful methodology for identifying differentially expressed genes. However, when thousands of genes in a microarray data set are evaluated simultaneously by fold changes and significance tests, the probability of detecting false positives rises sharply. In this first microarray study of brachial plexus injury, we applied and compared the performance of two recently proposed algorithms for tackling this multiple testing problem, Significance Analysis of Microarrays (SAM) and Westfall and Young step down adjusted p values, as well as t-statistics and Welch statistics, in specifying differential gene expression under different biological states.

RESULTS

Using SAM based on t statistics, we identified 73 significant genes, which fall into different functional categories, such as cytokines / neurotrophin, myelin function and signal transduction. Interestingly, all but one gene were down-regulated in the patients. Using Welch statistics in conjunction with SAM, we identified an additional set of up-regulated genes, several of which are engaged in transcription and translation regulation. In contrast, the Westfall and Young algorithm identified only one gene using a conventional significance level of 0.05.

CONCLUSION

In coping with multiple testing problems, Family-wise type I error rate (FWER) and false discovery rate (FDR) are different expressions of Type I error rates. The Westfall and Young algorithm controls FWER. In the context of this microarray study, it is, seemingly, too conservative. In contrast, SAM, by controlling FDR, provides a promising alternative. In this instance, genes selected by SAM were shown to be biologically meaningful.

Microarray gene expression profiles in dilated and hypertrophic cardiomyopathic end-stage heart failure

Despite similar clinical endpoints, heart failure resulting from dilated cardiomyopathy (DCM) or hypertrophic cardiomyopathy (HCM) appears to develop through different remodeling and molecular pathways. Current understanding of heart failure has been facilitated by microarray technology. We constructed an in-house spotted cDNA microarray using 10,272 unique clones from various cardiovascular cDNA libraries sequenced and annotated in our laboratory. RNA samples were obtained from left ventricular tissues of precardiac transplantation DCM and HCM patients and were hybridized against normal adult heart reference RNA. After filtering, differentially expressed genes were determined using novel analyzing software. We demonstrated that normalization for cDNA microarray data is slide-dependent and nonlinear. The feasibility of this model was validated by quantitative real-time reverse transcription-PCR, and the accuracy rate depended on the fold change and statistical significance level. Our results showed that 192 genes were highly expressed in both DCM and HCM (e.g., atrial natriuretic peptide, CD59, decorin, elongation factor 2, and heat shock protein 90), and 51 genes were downregulated in both conditions (e.g., elastin, sarcoplasmic/endoplasmic reticulum Ca2+-ATPase). We also identified several genes differentially expressed between DCM and HCM (e.g., alphaB-crystallin, antagonizer of myc transcriptional activity, beta-dystrobrevin, calsequestrin, lipocortin, and lumican). Microarray technology provides us with a genomic approach to explore the genetic markers and molecular mechanisms leading to heart failure.

Insights from gene arrays on the development and growth regulation of uterine leiomyomata

OBJECTIVE

To use microarray analysis as an unbiased approach to identify genes involved in the induction and growth of uterine leiomyomata.

DESIGN

Screen by arrays for up to 12,000 genes in leiomyoma (L) and control myometrium (M) from nine patients.

SETTING

University research laboratories.

PATIENT(S)

Nine patients in the follicular and luteal phases of the menstrual cycle.

INTERVENTION(S)

mRNA from L and M was converted to biotin-labeled cRNA and hybridized to cDNA oligonucleotide sequences on the arrays.

MAIN OUTCOME MEASURE(S)

Greater than two-fold change in gene expression between leiomyoma and matched myometrium.

RESULT(S)

Prominent among the 67 genes overexpressed in L relative to M were dlk or Pref-1, doublecortin, JM27, ionotropic glutamate receptor subunit 2, apolipoprotein E3, IGF2, semaphorin F, myelin proteolipid protein, MEST, frizzled, CRABP II, stromelysin-3, and TGFbeta3. The genes dlk, IGF2, and MEST are paternally expressed imprinted genes, and the others are involved in tissue differentiation and growth. Prominent among the 78 genes down-regulated in L relative to M were alcohol dehydrogenases 1alpha-gamma, tryptase, dermatopontin, thrombospondin, coxsackievirus receptor, nur77, and c-kit.

CONCLUSION(S)

Arrays offer large-scale screening of mRNA expression, which will help us differentiate between the genes and metabolic pathways necessary for leiomyoma growth and those regulating myometrial contractions.

Microarray technology and gene expression analysis for the study of angiogenesis

Angiogenesis plays a major role in multiple disease processes including cancer, and new agents that modulate angiogenesis are rapidly entering clinical trials. The understanding of the biological mechanisms and downstream effects for many of these agents is poorly understood. It is therefore important that methods evolve to understand how an agent regulates angiogenesis, in order to promote a higher percentage of successful drug candidates. With the emergence of microarray technology for the evaluation of gene expression, researchers have a powerful tool for dissecting the biological mechanisms of angiogenesis. However, huge data sets and complex statistics pose a hurdle for the investigator to obtain useful and meaningful data. To eliminate problems in data analysis, proper design and planning prior to performing a microarray experiment is crucial to making valid conclusions. This review will discuss the critical factors in designing, performing and analysing microarray experiments, and the utility of various models of angiogenesis for microarray analysis.

Gene profiling techniques and their application in angiogenesis and vascular development

The analysis of gene expression in specific tissues and physiological processes has evolved over the last 20 years from the painstaking identification of selected genes to the relatively efficient and open-ended surveying of potentially all genes expressed in a tissue. Current art for gene discovery includes the use of large-scale arrays of cDNA sequences or oligonucleotides, and molecular 'tagging' techniques such as GeneCalling and SAGE. Common to each of these techniques is a reliance on the increasingly comprehensive databases of human and mouse EST and full-length gene sequences. Early efforts to characterize candidate genes were limited by their narrow scope, while current efforts are confounded by the enormous volume of data returned. Sophisticated software tools are an integral part of the analysis, helping to organize information into coherent groups with temporal or functional similarity. These techniques, in conjunction with the continued analysis of human genetic syndromes, transgenic, and knockout mice, have driven genetic analysis of angiogenesis and vascular development from describing which individual genes are involved to defining the outlines of regulatory networks.