Clustering Algorithms: On Learning,Validation, Performance,and Applications to Genomics

The development of microarray technology has enabled scientists to measure the expression of thousands of genes simultaneously, resulting in a surge of interest in several disciplines throughout biology and medicine. While data clustering has been used for decades in image processing and pattern recognition, in recent years it has joined this wave of activity as a popular technique to analyze microarrays. To illustrate its application to genomics, clustering applied to genes from a set of microarray data groups together those genes whose expression levels exhibit similar behavior throughout the samples, and when applied to samples it offers the potential to discriminate pathologies based on their differential patterns of gene expression. Although clustering has now been used for many years in the context of gene expression microarrays, it has remained highly problematic. The choice of a clustering algorithm and validation index is not a trivial one, more so when applying them to high throughput biological or medical data. Factors to consider when choosing an algorithm include the nature of the application, the characteristics of the objects to be analyzed, the expected number and shape of the clusters, and the complexity of the problem versus computational power available. In some cases a very simple algorithm may be appropriate to tackle a problem, but many situations may require a more complex and powerful algorithm better suited for the job at hand. In this paper, we will cover the theoretical aspects of clustering, including error and learning, followed by an overview of popular clustering algorithms and classical validation indices. We also discuss the relative performance of these algorithms and indices and conclude with examples of the application of clustering to computational biology.Key Words: Clustering, genomics, profiling, microarray, validation index.     

Gene expression profiling and molecular characterization of antimony resistance in Leishmania amazonensis

Background

Drug resistance is a major problem in leishmaniasis chemotherapy. RNA expression profiling using DNA microarrays is a suitable approach to study simultaneous events leading to a drug-resistance phenotype. Genomic analysis has been performed primarily with Old World Leishmania species and here we investigate molecular alterations in antimony resistance in the New World species L. amazonensis.

Methods/Principal Findings

We selected populations of L. amazonensis promastigotes for resistance to antimony by step-wise drug pressure. Gene expression of highly resistant mutants was studied using DNA microarrays. RNA expression profiling of antimony-resistant L. amazonensis revealed the overexpression of genes involved in drug resistance including the ABC transporter MRPA and several genes related to thiol metabolism. The MRPA overexpression was validated by quantitative real-time RT-PCR and further analysis revealed that this increased expression was correlated to gene amplification as part of extrachromosomal linear amplicons in some mutants and as part of supernumerary chromosomes in other mutants. The expression of several other genes encoding hypothetical proteins but also nucleobase and glucose transporter encoding genes were found to be modulated.

Conclusions/Significance

Mechanisms classically found in Old World antimony resistant Leishmania were also highlighted in New World antimony-resistant L. amazonensis. These studies were useful to the identification of resistance molecular markers.     

DNA microarrays: raising the profile

Expression profiling using DNA microarrays is starting to come of age. The past year has seen significant advances in the number, scope and quality of studies that incorporate expression profiling experiments. Attention is starting to move on from making DNA microarrays to appropriate experimental design and sophisticated data analysis techniques.     

Protein microarrays and novel detection platforms

The field of proteomics has undergone rapid advancements over the last decade and protein microarrays have emerged as a promising technological platform for the challenging task of studying complex proteomes. This gel-free approach has found an increasing number of applications due to its ability to rapidly and efficiently study thousands of proteins simultaneously. Different protein microarrays, including capture arrays, reverse-phase arrays, tissue microarrays, lectin microarrays and cell-free expression microarrays, have emerged, which have demonstrated numerous applications for proteomics studies including biomarker discovery, protein interaction studies, enzyme-substrate profiling, immunological profiling and vaccine development, among many others. The need to detect extremely low-abundance proteins in complex mixtures has provided motivation for the development of sensitive, real-time and multiplexed detection platforms. Conventional label-based approaches like fluorescence, chemiluminescence and use of radioactive isotopes have witnessed substantial advancements, with techniques like quantum dots, gold nanoparticles, dye-doped nanoparticles and several bead-based methods now being employed for protein microarray studies. In order to overcome the limitations posed by label-based technologies, several label-free approaches like surface plasmon resonance, carbon nanotubes and nanowires, and microcantilevers, among others, have also advanced in recent years, and these methods detect the query molecule itself. The scope of this article is to outline the protein microarray techniques that are currently being used for analytical and function-based proteomics and to provide a detailed analysis of the key technological advances and applications of various detection systems that are commonly used with microarrays.     

Genome-directed primers for selective labeling of bacterial transcripts for DNA microarray analysis

DNA microarrays have the ability to analyze the expression of thousands of the same set of genes under at least two different experimental conditions. However, DNA microarrays require substantial amounts of RNA to generate the probes, especially when bacterial RNA is used for hybridization (50 microg of bacterial total RNA contains approximately 2 microg of mRNA). We have developed a computer-based algorithm for prediction of the minimal number of primers to specifically anneal to all genes in a given genome. The algorithm predicts, for example, that 37 oligonucleotides should prime all genes in the Mycobacterium tuberculosis genome. We tested the usefulness of the genome-directed primers (GDPs) in comparison to random primers for gene expression profiling using DNA microarrays. Both types of primers were used to generate fluorescent-labeled probes and to hybridize to an array of 960 mycobacterial genes. Compared to random-primer probes, the GDP probes were more sensitive and more specific, especially when mammalian RNA samples were spiked with mycobacterial RNA. The GDPs were used for gene expression profiling of mycobacterial cultures grown to early log or stationary growth phases. This approach could be useful for accurate genome-wide expression analysis, especially for in vivo gene expression profiling, as well as directed amplification of sequenced genomes.     

3′-End Sequencing for Expression Quantification (3SEQ) from Archival Tumor Samples

Gene expression microarrays are the most widely used technique for genome-wide expression profiling. However, microarrays do not perform well on formalin fixed paraffin embedded tissue (FFPET). Consequently, microarrays cannot be effectively utilized to perform gene expression profiling on the vast majority of archival tumor samples. To address this limitation of gene expression microarrays, we designed a novel procedure (3′-end sequencing for expression quantification (3SEQ)) for gene expression profiling from FFPET using next-generation sequencing. We performed gene expression profiling by 3SEQ and microarray on both frozen tissue and FFPET from two soft tissue tumors (desmoid type fibromatosis (DTF) and solitary fibrous tumor (SFT)) (total n = 23 samples, which were each profiled by at least one of the four platform-tissue preparation combinations). Analysis of 3SEQ data revealed many genes differentially expressed between the tumor types (FDR<0.01) on both the frozen tissue (∼9.6K genes) and FFPET (∼8.1K genes). Analysis of microarray data from frozen tissue revealed fewer differentially expressed genes (∼4.64K), and analysis of microarray data on FFPET revealed very few (69) differentially expressed genes. Functional gene set analysis of 3SEQ data from both frozen tissue and FFPET identified biological pathways known to be important in DTF and SFT pathogenesis and suggested several additional candidate oncogenic pathways in these tumors. These findings demonstrate that 3SEQ is an effective technique for gene expression profiling from archival tumor samples and may facilitate significant advances in translational cancer research.     

A Rapid Ribosome Profiling Method Elucidates Chloroplast Ribosome Behavior in Vivo

The profiling of ribosome footprints by deep sequencing has revolutionized the analysis of translation by mapping ribosomes with high resolution on a genome-wide scale. We present a variation on this approach that offers a rapid and cost-effective alternative for the genome-wide profiling of chloroplast ribosomes. Ribosome footprints from leaf tissue are hybridized to oligonucleotide tiling microarrays of the plastid ORFeome and report the abundance and translational status of every chloroplast mRNA. Each assay replaces several time-consuming traditional methods while also providing information that was previously inaccessible. To illustrate the utility of the approach, we show that it detects known defects in chloroplast gene expression in several nuclear mutants of maize (Zea mays) and that it reveals previously unsuspected defects. Furthermore, it provided firm answers to several lingering questions in chloroplast gene expression: (1) the overlapping atpB/atpE open reading frames, whose translation had been proposed to be coupled, are translated independently in vivo; (2) splicing is not a prerequisite for translation initiation on an intron-containing chloroplast RNA; and (3) a feedback control mechanism that links the synthesis of ATP synthase subunits in Chlamydomonas reinhardtii does not exist in maize. An analogous approach is likely to be useful for studies of mitochondrial gene expression.     

The Maestro (Mro) Gene Is Dispensable for Normal Sexual Development and Fertility in Mice

The mammalian gonad arises as a bipotential primordium from which a testis or ovary develops depending on the chromosomal sex of the individual. We have previously used DNA microarrays to screen for novel genes controlling the developmental fate of the indifferent embryonic mouse gonad. Maestro (Mro), which encodes a HEAT-repeat protein, was originally identified as a gene exhibiting sexually dimorphic expression during mouse gonad development. Wholemount in situ hybridisation analysis revealed Mro to be expressed in the embryonic male gonad from approximately 11.5 days post coitum, prior to overt sexual differentiation. No significant expression was detected in female gonads at the same developmental stage. In order to address its physiological function, we have generated mice lacking Maestro using gene targeting. Male and female mice homozygous for a Mro null allele are viable and fertile. We examined gonad development in homozygous male embryos in detail and observed no differences when compared to wild-type controls. Immunohistochemical analysis of homozygous mutant testes of adult mice revealed no overt abnormalities. Expression profiling using DNA microarrays also indicated no significant differences between homozygote embryonic male gonads and controls. We conclude that Maestro is dispensable for normal male sexual development and fertility in laboratory mice; however, the Mro locus itself does have utility as a site for insertion of transgenes for future studies in the fields of sexual development and Sertoli cell function.     

Biologically meaningful expression profiling across species using heterologous hybridization to a cDNA microarray

Background  

Unravelling the path from genotype to phenotype, as it is influenced by an organism's environment, is one of the central goals in biology. Gene expression profiling by means of microarrays has become very prominent in this endeavour, although resources exist only for relatively few model systems. As genomics has matured into a comparative research program, expression profiling now also provides a powerful tool for non-traditional model systems to elucidate the molecular basis of complex traits.     

Use of arrays to investigate the contribution of ATP-binding cassette transporters to drug resistance in cancer chemotherapy and prediction of chemosensitivity

Multidrug resistance （MDR） is a major problem in cancer chemotherapy. One of the best known mechanisms of MDR is the elevated expression of ATP-binding cassette （ABC） transporters. While some members of human ABC transporters have been shown to cause drug resistance with elevated expression, it is not yet known whether the over-expression of other members could also contribute to drug resistance in many model cancer cell lines and clinics. The recent development ofmicroarrays and quantitative PCR arrays for expression profiling analysis of ABC transporters has helped address these issues. In this article, various arrays with limited or full list of ABC transporter genes and their use in identifying ABC transporter genes in drug resistance and chemo-sensitivity prediction will be reviewed.     

News in Brief

Protein microarrays are versatile tools for parallel, miniaturized screening of binding events involving large numbers of immobilized proteins in a time- and cost-effective manner. They are increasingly applied for high-throughput protein analyses in many research areas, such as protein interactions, expression profiling and target discovery. While conventionally made by the spotting of purified proteins, recent advances in technology have made it possible to produce protein microarrays through in situ cell-free synthesis directly from corresponding DNA arrays. This article reviews recent developments in the generation of protein microarrays and their applications in proteomics and diagnostics.     

Protein expression profiling arrays: tools for the multiplexed high-throughput analysis of proteins

The completion of the human genome sequence has led to a rapid increase in genetic information. The invention of DNA microarrays, which allow for the parallel measurement of thousands of genes on the level of mRNA, has enabled scientists to take a more global view of biological systems. Protein microarrays have a big potential to increase the throughput of proteomic research. Microarrays of antibodies can simultaneously measure the concentration of a multitude of target proteins in a very short period of time. The ability of protein microarrays to increase the quantity of data points in small biological samples on the protein level will have a major impact on basic biological research as well as on the discovery of new drug targets and diagnostic markers. This review highlights the current status of protein expression profiling arrays, their development, applications and limitations.