This collection of chapters in the field of bioinformatics focuses on quantitative analysis. A number of examples accompany the methodology, so readers can easily understand how to apply the methods to solve problems.
The first two chapters address data clustering. Modern biomedical instruments can generate large amounts of data. One task is to find variables that can distinguish phenotypes. The popular unsupervised approach involves the use of clustering techniques, such as biclustering, which attempts to simultaneously partition phenotypes and variables. For example, in microarray analysis, we want to identify up-regulated genes. Another important example is the need to cluster time series in biomedical data analysis. Since biological systems are dynamic, we need to collect the time series data to capture such dynamics. When performing time series clustering, we can partition variables with similar temporal behavior into groups.
Chapter 3 presents a supervised technique using a support vector machine (SVM). In bioinformatics, we sometimes use known samples to create a classifier, which can be used to classify the unknown samples. SVMs are developed to create classifiers that can find a hyperplane in the sample space to separate different classes of samples. Since biological systems usually possess nonlinearity, a nonlinear SMV was developed to design better classifiers.
Chapter 4 uses Alzheimer’s disease as an example to illustrate the construction of predictive models. In biomedicine, predicting disease progression is an important topic because it might allow physicians to foresee how patients will progress over time. The techniques of predictive models can be extended to investigate other diseases as well.
Chapter 5 focuses on solving missing data problems. During biomedical data collection, we frequently encounter missing data. For instance, some patients enrolled in a study may miss an appointment, so their data cannot be collected. Developing a way to estimate the values for the missing data can remedy the incompleteness and lay the foundation for a comprehensive analysis.
Chapters 6 and 7 address biomedical image informatics. Chapter 6 introduces cardiac image processing, and chapter 7 discusses functional magnetic resonance imaging (fMRI). The major tasks involve removing noise and segmenting the regions of interest.
Chapters 8 and 9 discuss biomedical signal processing. Biomedical signal measurements are frequently corrupted by noise. We need methods to extract crucial patterns to better understand the underlying biological mechanisms. These chapters present different approaches to the analysis of biomedical signals, using neural signal processing as an example to illustrate the applications.
The last chapter introduces graph theory and its application in the analysis of neural networks in the brain. Using graph theoretical techniques can model neural connectivity and reverse-engineer how neurons communicate with each other.
In summary, this book provides a comprehensive introduction to the quantitative analysis of biomedical data. It is a great handbook for beginners entering the field of bioinformatics.