Clustering in Bioinformatics and Drug Discovery

Biological and pharmaceutical datasets can be vast, varied, and difficult to interpret. Clustering in Bioinformatics and Drug Discovery examines how cluster analysis can help organize those data—and, just as importantly, how to assess whether a clustering method suits the problem at hand.

John D. MacCuish and Norah E. MacCuish build from data representation and similarity measures toward a broad selection of clustering approaches, with applications ranging from compound libraries and screening results to gene-expression and protein-reaction data.

Start with the data, not just the algorithm

Early chapters introduce data types, normalization, transformations, similarity and proximity measures, matrices, dimensionality methods, and graph theory. This grounding helps readers see why the choice of representation and measure can shape the clusters an algorithm produces.

A wide toolkit for clustering

The book surveys partitional, hierarchical, sampling, hybrid, overlapping, and self-organizing approaches. Its algorithm coverage includes methods such as K-means, Jarvis–Patrick, spectral clustering, self-organizing maps, leader algorithms, and Taylor–Butina. Later material considers asymmetry, ambiguity, validation, visualization, and large-scale or parallel algorithms.

Applications in bioinformatics and drug discovery

Examples and applications connect the methods to practical research questions, including clustering large combinatorial libraries, compound diversity and acquisition, high-throughput screening, lead hopping, gene-expression toxicity studies, and protein reaction data. The focus on domain-specific issues makes this more than a general survey of clustering techniques.

For readers who need to judge their results

This title will interest computational biology and bioinformatics readers, as well as researchers and practitioners in cheminformatics, statistics, and pharmaceutical discovery. Its attention to validation and clustering ambiguity is especially relevant when a result needs interpretation—not merely a cluster assignment.

Three appendices offer primers on matrix algebra, probability theory, and number theory, supporting readers as they work through the book’s mathematical and computational ideas.