Introduction to Bioinformatics

Citation

Lesk, A. M. (2019). Introduction to Bioinformatics (5th ed.). Oxford University Press.

Chapter Summary

The fifth edition runs to about 400 pages in ten chapters and a short closing section of conclusions. Chapters carry exercises and problems, and lists of recommended readings; Oxford's companion website adds answers to the exercises, hints to the problems, "weblems" (web-based problems), the book's Perl programs and a list of web links. The publisher lists new material on next-generation sequencing, function prediction, sequence assembly, epigenomics, the bioinformatics of gene editing and the effects of single nucleotide variants.

Chapter 1: Introduction

At about 47 pages, a substantial opening. The exercises show it reviewing the chemistry of DNA and RNA, the genetic code and translation, and introducing graphs (nodes and edges) as a way to represent relationships. Two case studies use short Perl programs, one translating a DNA sequence into protein and one assembling overlapping fragments, and a weblem asks students to explain the discovery of CRISPR/Cas.

Chapter 2: From genetics to genomes

Moves from classical genetics to whole genomes: restriction maps and gel electrophoresis, genome sizes across the tree of life, sequencing, contigs and coverage, and the string and graph methods used in assembly, including suffix trees and Eulerian paths. The exercises also apply DNA evidence to identification and kinship questions, such as paternity, sex determination and the identification of Richard III.

Chapter 3: The panorama of life

A tour of sequenced genomes and what they show, with tables of completed eukaryotic genomes and the status of genome projects as of March 2018. Exercises and weblems cover patterns of inheritance of genetic diseases, mitochondrial inheritance, operons, horizontal gene transfer, the duplications that produced the human globin genes, and the classification of protein function.

Chapter 4: Alignments and phylogenetic trees

The core methods chapter and one of the longest in the book. It covers dotplots (with a Perl program in Box 4.1), substitution matrices such as PAM and BLOSUM, dynamic-programming alignment, database searching with BLAST and PSI-BLAST, multiple sequence alignment and profiles, and the construction of phylogenetic trees from distance matrices. A sequence of weblems walks up the tree of life from primates to deuterostomes.

Chapter 5: Structural bioinformatics and drug discovery

Introduces protein structure: the forces that stabilise it, secondary structure and common folding motifs such as hairpins, beta-alpha-beta units and TIM barrels, orthologues and paralogues, coiled-coil prediction, and the use of the Protein Data Bank. Exercises then turn to drug design, ligand binding and the thermodynamics of binding.

Chapter 6: Scientific publications and archives: media, content, access, and presentation

A chapter on how biological knowledge is stored and found: the scientific literature and PubMed, open access and copyright, databases such as the Protein Data Bank, and the retrieval of information from text, including recognising gene and protein names and the problem of ambiguity in natural language.

Chapter 7: Artificial intelligence and machine learning

Introduces classification and its evaluation (true and false positives, ROC curves and the area under them, training sets), dimensionality reduction with principal component analysis, neural networks built from weights and thresholds, including the classic limit that a network without hidden layers cannot compute an exclusive or, and mapping data into higher-dimensional feature spaces. A case study distinguishes protein-coding from non-coding RNA transcripts, and weblems apply the methods to protein structure classifications.

Chapter 8: Introduction to systems biology

Returns to graphs as models of biological networks: paths and shortest routes (worked through with the London Underground), degree distributions, and scale-free networks and their growth. The problems also bring in ideas from information theory, such as how many yes-or-no questions are needed to distinguish items, data compression and the Burrows-Wheeler transform.

Chapter 9: Metabolic pathways

Covers enzyme kinetics, including the Michaelis-Menten relation, the naming of enzymes by Enzyme Commission and Gene Ontology identifiers, specific biosynthetic pathways, and the comparison of pathways between species.

Chapter 10: Control of organization and organization of control

The final chapter treats regulation and the organisation of the cell: experimental methods for detecting interactions, such as phage display, two-hybrid screens and structure determination, the energetics of interactions, and the wiring of transcriptional regulatory networks, including recurring connection patterns between regulators and target genes. A brief section of conclusions closes the book.

Key Concepts

Chapters 1–3: Biological and genomic background

• Genetic code and translation: the mapping from codons to amino acids, implemented as the first program in the book.
• Genome assembly: rebuilding a genome from overlapping reads, with coverage and contigs as measures of progress and suffix trees and Eulerian paths as tools.
• Comparative genomics: what the genomes of different organisms show about gene content, gene duplication and horizontal transfer.

Chapters 4–5: Sequences and structures

• Substitution matrices: tables such as PAM and BLOSUM that score how likely one amino acid is to replace another.
• Dynamic programming: the method that finds optimal alignments by building on optimal solutions to smaller problems.
• Database searching: BLAST and the profile-based PSI-BLAST for finding distant homologues.
• Protein structure hierarchy: secondary structures and recurring motifs that combine into folds.

Chapters 6–7: Information and learning

• Information retrieval: finding relevant records in literature and archives, and the difficulty of extracting meaning from text.
• Classifier evaluation: sensitivity, specificity and ROC curves for judging prediction methods.
• Neural networks: layers of weighted units with thresholds that learn to classify data.

Chapters 8–10: Systems

• Biological networks as graphs: nodes, edges, paths and degree distributions applied to metabolic and regulatory systems.
• Enzyme kinetics: the dependence of reaction rate on substrate concentration that underlies pathway models.
• Regulatory networks: the control circuits that coordinate gene expression and cell organisation.

Critical Analysis

Strengths

• Written for biologists. The publisher states that the book assumes no prior knowledge and is written primarily for a biological audience without detailed knowledge of programming. The opening chapters review molecular biology and genetics before methods arrive.
• Plenty of practice, with support. The publisher describes frequent examples, self-test questions, exercises and problems. The companion site publishes answers to the exercises and hints to the problems, so self-study is practical, and the weblems send students to real databases and servers.
• Breadth with a light computational thread. The book covers genomes, alignment, phylogeny, structure, literature databases, machine learning, networks, pathways and regulation. Short Perl programs in the text show what the computations look like without making programming a prerequisite. Endorsements quoted by the publisher, from university lecturers, praise its clear writing and its mix of narrative and diagrams.
• Applications beyond the laboratory. Forensic identification, kinship and genetic disease recur in the exercises, which the publisher describes as part of the book's scope (medicine, forensic science, anthropology).

Limitations

• Programming is shown, not taught. The programs are in Perl, now less used in bioinformatics than Python; a few weblems ask for Python programs, but the book does not teach a language. Students who need to write analysis code need a separate course or a book such as Compeau and Pevzner's.
• Methods at an overview level. Each chapter covers a broad area in 20–55 pages, so algorithms are explained but not developed in depth, and statistical treatment of scores and of high-throughput data is limited.
• Dated in fast-moving areas. The 2019 edition reports genome-project status as of March 2018 and machine-learning methods of that period; newer sequencing technologies, databases and deep-learning tools are not covered.

Where It Fits in the Course

This is a gentle entry point to the course. Compeau and Pevzner's Bioinformatics Algorithms is also written for newcomers, but it starts from biological questions and develops the algorithms through programming challenges; students who want to write code should pair the two, reading Lesk's chapters 2 and 4 alongside its chapters on assembly and alignment. The Baxevanis, Bader and Wishart Bioinformatics is an edited textbook by specialists that goes further into the practical use of databases and analysis tools, from genome annotation to expression data and metagenomics; it is the natural next book once Lesk's overview is in place.

Zvelebil and Baum's Understanding Bioinformatics gives a longer and more mathematical account of sequence, structure and systems methods, suited to students who want the derivations behind Lesk's descriptions. Uri Alon's An Introduction to Systems Biology is a quantitative, differential-equation treatment of network motifs and design principles; Lesk's chapters 8–10 give the qualitative background for it, but Alon requires calculus that Lesk does not.

How to Use This Book

The book suits first- or second-year biology students, and biomedical or other life-science students, who need a first course in bioinformatics; introductory molecular biology is enough background, and no programming is required. Students from computing will find the biology review in chapters 1–3 useful and the computational parts light.

A suggested plan:

• Core: chapters 1, 2 and 4. Together they give the biological background, sequencing and assembly, and the alignment and search methods used throughout bioinformatics. Work the exercises in chapter 4 by hand before running real searches.
• Second layer: chapters 3 and 5, for comparative genomics and protein structure; chapter 5 is essential for students heading towards structural biology or pharmacology.
• Selective: chapters 6 and 7 can be read according to need; chapter 7 is the place to learn how prediction methods are judged. Chapters 8–10 form a short introduction to systems biology and can be taken as a unit at the end of a course.

Use the companion website actively: check exercise answers there, use the problem hints before giving up, and do several weblems per chapter, since they connect the text to the databases and servers used in practice. Run the Perl programs, or rewrite them in Python, to see the algorithms work. Because the field moves quickly, check the current versions of the databases and tools that the book mentions.