Abstract
Computational models have been playing a significant role for the computer-based analysis of biological and biomedical data. Given the recent availability of genomic sequences and microarray gene expression, and proteomic data, there is an increasing demand for developing and applying advanced computational techniques for exploring these types of data such as: functional interpretation of gene expression data, deciphering of how genes, and proteins work together in pathways and networks, extracting and analysing phenotypic features of mitotic cells for high throughput screening of novel anti-mitotic drugs. Successful applications of advanced computational algorithms to solving modern life-science problems will make significant impacts on several important and promising issues related to genomic medicine, molecular imaging, and the scientific knowledge of the genetic basis of diseases. This chapter reviews the fusion of engineering, computer science, and information sciences with biology and medicine to address some latest technical developments in the computational analyses of modern biological data: microarray gene expression data, mass spectrometry data, and bioimaging.
TopMicroarray Gene Expression Data
Microarrays are a relatively new biotechnology that provides novel insights into gene expression and gene regulation (Brazma and Vilo, 2000; Whitchurch, 2002; Zhang at al, 2002; Pham et al, 2006a). Microarray technology has been applied in diverse areas ranging from genetics and drug discovery to disciplines such as virology, microbiology, immunology, endocrinology, and neurobiology. Microarray-based methods are the most widely used technology for large-scale analysis of gene expression because they allow simultaneous study of mRNA abundance for thousands of genes in a single experiment (Kellum and Liu, 2003). The generation of DNA microarray image spots involves the hybridization of two probes labelled with a fluorescent red dye or a fluorescent green dye. The relative image intensity values of the red dye and the green dye on a particular spot of the arrays indicate the expression ratio for the corresponding gene of the two samples from which the mRNAs have been extracted. Thus, robust image processing of microarray spots plays an important role in microarray technology (Nagarajan, 2003; Liew at al, 2003; Lukac et al, 2004).
DNA microarray data consists of a large number of genes and a relatively small number of experimental samples. The number of genes on an array is in the order of thousands, and because this far exceeds the number of samples, dimension reduction is needed to allow efficient analysis of data classification techniques. Many statistical and machine-learning techniques based different computational methodologies have been applied for cancer classification in microarray experiments. These techniques include linear discriminant analysis, k-nearest neighbor algorithms, Bayes classifiers, decision trees, neural networks, and support vector machines (Dudoit and Fridlyand, 2003; Golub et al, 1999; Guyon et al, 2002). Nevertheless, common tasks of most classifiers are to perform feature selection and decision logic.
Based on the motivation that conventional statistical methods for pattern classification break down when there are more variables (genes) than there are samples, Nguyen and Rocke (2002) proposed a partial least-squares method for classifying human tumor samples using microarray gene expression data. Zhou et al. (2004) proposed a Bayesian approach for selecting the strongest genes based on microarray gene expression data and the logistic regression model for classifying and predicting cancer genes. Yeung et al. (2005) reported that conventional methods for gene selection and classification do not take into account model uncertainty and use a single set of selected genes for prediction, and introduced a Bayesian model averaging method, which considers the uncertainty by averaging over multiple sets of overlapping relevant genes. Furey et al. (2000) applied support vector machines for the classification of cancer tissue samples or cell types using microarrays. Lee et al. (2003) proposed a Bayesian model for gene selection for cancer classification using microarray data. Statnikov et al. (2005) carried out a comprehensive evaluation of classification methods for cancer diagnosis based on microarray gene expression data.
Recently Pham et al. (2006b) carried out cancer classification by transforming microarray data into spectral vectors. The same authors used the spectral difference or spectral distortion between the pair of spectra for pattern comparison, which appears to be a potential approach for the cancer classification using microarray gene expression data.
Key Terms in this Chapter
Time-Lapse Microcopy Imaging: Microscopy imaging that captures images of dynamic events at predetermined time intervals.
Naive Bayes Classifier: A c lassification technique that is based on the so-called Bayesian theorem.
Feature Extraction: Extraction of representative properties of an object for the purpose of classification.
Support Vector Machines: machine learning algorithms that map input vectors to a higher dimensional space where a maximal separating hyperplane is constructed.
Wavelet Transform: The representation of a signal in terms of scaled and translated copies of a finite length or fast decaying oscillating waveform.
Linear Predictive Coding: An encoding method that allows the prediction of the value of the signal at each sample as a linear combination of the past values of the signal.
Proteomics: The study of the structure and function of proteins,
Geostatistics: Applied statistics of spatially correlated data
High Content Screening: A high throughput platform for understanding the functions of genes, RNA, proteins, and other cellular constituents at the level of the living cell.
k-Nearest Neighbor Algorithms: Methods for classifying objects based on closest training samples in the feature space.
Genetic Algorithms: Biologically inspired optimization methods.
Spectral Distortion: A measure of mismatch between two signals based on their spectral properties.
Artificial Neural Networks: Machine learning methods consisting of interconnecting artificial neurons that simulate the properties of biological neural networks.
Feature Reduction: Compression of the feature space of an object.
Cluster Analysis: Methods for grouping objects of similar kind into respective categories.
Microarray Gene Expression Data: Modern biotechnological data generated for studying the interaction of large numbers of genes and how a cell’s regulatory networks control genes simultaneously.
Biomarker Discovery: Discovery of molecular parameters associated with the presence and severity of specific disease states.
Decision Trees: Predictive models that map the observations about an event to infer about its target value.
Mass Spectrometry Data: A dataset that consists of relative intensities a chromatographic retention time and the ratios of molecular mass over charge. The mass spectrum for a sample is a function of the molecules and used to test for presence or absence of one or more molecules which may relate to a diseased state or a cell type.