Microarrays for an Integrative Genomics (Computational Molecular Biology) [Hardcover]

Isaac S. Kohane (Author), Alvin Kho (Author), Atul J. Butte (Author)

Book Description

Series: Computational Molecular Biology | Publication Date: August 21, 2002

Functional genomics--the deconstruction of the genome to determine the biological function of genes and gene interactions--is one of the most fruitful new areas of biology. The growing use of DNA microarrays allows researchers to assess the expression of tens of thousands of genes at a time. This quantitative change has led to qualitative progress in our ability to understand regulatory processes at the cellular level.This book provides a systematic introduction to the use of DNA microarrays as an investigative tool for functional genomics. The presentation is appropriate for readers from biology or bioinformatics. After presenting a framework for the design of microarray-driven functional genomics experiments, the book discusses the foundations for analyzing microarray data sets, genomic data-mining, the creation of standardized nomenclature and data models, clinical applications of functional genomics research, and the future of functional genomics.

Editorial Reviews

Review

"This book should be required reading for any (future and present) post-genomic researcher..." Sean D. Mooney Briefings in Bioinformatics

About the Author

Isaac S. Kohane is Director of the Children's Hospital Informatics Program, Associate Professor of Pediatrics at Harvard Medical School, and an Attending Physician in Endocrinology at Children's Hospital, Boston.

Alvin Kho is Research Fellow in Medicine at Children's Hospital, Boston.

Atul J. Butte is an Assistant Professor of Medicine (Medical Informatics) and Pediatrics at the Stanford University School of Medicine.

Product Details

• Hardcover: 326 pages
• Publisher: A Bradford Book; 1 edition (August 21, 2002)
• Language: English
• ISBN-10: 026211271X
• ISBN-13: 978-0262112710
• Product Dimensions: 9.4 x 7.2 x 0.8 inches
• Shipping Weight: 1.5 pounds (View shipping rates and policies)
• Average Customer Review: 4.4 out of 5 stars  See all reviews (9 customer reviews)
• Amazon Best Sellers Rank: #2,077,672 in Books (See Top 100 in Books)

More About the Author

Discover books, learn about writers, read author blogs, and more.

Customer Reviews

Most Helpful Customer Reviews

4 of 4 people found the following review helpful:

4.0 out of 5 stars A helpful and informative overview, January 4, 2006

By 

Dr. Lee D. Carlson (Baltimore, Maryland USA) - See all my reviews
(VINE VOICE)    (HALL OF FAME REVIEWER)    (REAL NAME)   

Amazon Verified Purchase

This review is from: Microarrays for an Integrative Genomics (Computational Molecular Biology) (Paperback)

The authors of this book are very excited about the prospects of the field of functional genomics and DNA microarray technology. Their optimism however is tempered by a large degree of caution, for they make it clear in the first few paragraphs of the book that expression profiling using microarrays is still in its infancy and that there have been exaggerated reports of its success. They wrote this book with the intent of giving the reader a more realistic view of microarray technology and have succeeded in their goal. They target the book specifically to experienced biologists and bioinformaticians with limited experience in using microarrays, and to students who are entering the field of bioinformatics. Most importantly, they emphasize that functional genomics is an experimental science, and that highly sophisticated algorithms from data mining or other areas of artificial intelligence will be of no assistance if the experimental information is not there in the first place. They do encourage however further development of these algorithms, in order to be able to extract the data as it becomes available, and as microarray technology itself matures. Even with the current technology, enormous amounts of data are generated, and if sense is to be made of this data, one will have to develop more effective algorithms than what are currently available.

To perform successful experiments, the authors describe a `functional genomics pipeline', and list the characteristics that it must have, consisting of both `wet' (laboratory) and `dry' (computational) steps. They devote a lot of space in the book describing how to develop an effective genomic experiment. Crucial to such investigations they say is a design that maximizes the possibility of observing relevant gene expression patterns, and the `experiment design space', which encapsulates all possible conditions that a particular biological system could be influenced by. Also important to the design is the `expression space', which is the collection of all potential expression values of all genes in a given genome. One could view the expression space as a vector space of high dimension, with each dimension corresponding to a single gene. Of great interest, and widely discussed in the general bioinformatics literature under the guise of the new field of `systems biology' is a subset of the expression space called the `transcriptome.' This subset models the expression of a cellular system under all stimuli. Considering that one might have to deal with 30,000 genes in the case of a human, the characterization of the transcriptome will be a formidable project. Interactions between the genes will complicate the analysis even further. The authors view each experiment as being an exploration of the space of all possible expression patterns, and describe good experimentation as being the `maximal exercise' of the genome. This consists of finding those correlations between the genes that have the greatest impact on the process under scrutiny.

A book on microarrays would not be complete if it did not discuss how they actually function. This is done in a fair detail in chapter three of the book. The authors do not favor a particular vendor but rather discuss what biological assumptions all microarray technology is based on. One of these assumptions is, as expected, that there is a direct connection between mRNA transcription and the protein translation associated with it.

In any laboratory experiment one has to deal with experimental uncertainty or "noise." This involves the influence of unknown external perturbations that result in variability in the outcomes of the experiment. As further evidence that the authors are careful experimenters, they discuss noise in detail, noting first that expression experiments deal with information that is both digital (DNA sequence information) and analog (mRNA expression levels). They distinguish between `intra-chip' noise, which arises when one probe feature influences another, improper scanning techniques, and manufacturing defects, and `inter-chip' noise, which arises from sample variation. Normalization issues are also discussed. Readers should take particular attention to the discussion on fold calculation and significance because of its connection with statistical analysis and because it sets the tone for the rest of the book. In particular, this discussion leads to the very important topic of dissimilarity and similarity measures. This part of the book is more sophisticated mathematically than what has been encountered so far, dealing for example with the concept of a metric space, which may appear to be somewhat abstract by readers who are not mathematically astute. Linear correlation and mutual information are two examples of metrics that are discussed.

Data mining is of course heavily discussed in the book, along with the new field of `ontological engineering' and how the latter is used functional genomics. Data mining is of course a vast field, but the authors give the reader a good taste of how some of its techniques can be applied to analyze microarray experiments. Both unsupervised and supervised learning is discussed, along with `self-organizing maps.' The authors end the book with their vision of future developments. Naturally they point to further refinements in microarray technology, the need for educating a new generation of bioinformaticists, and the push towards the development of new data mining algorithms. Certainly all of these are important, and one can expect other technological developments to occur in the coming years that may prove superior to microarrays in their application to functional genomics. In addition, and there are indications of this even at the present time, one can expect technologies that fully automate the study of gene expression. This includes the generation of hypotheses that characterize scientific investigation, the development and construction of the experiments themselves, and the analysis of the resulting data.

4 of 4 people found the following review helpful:

5.0 out of 5 stars Well written, June 28, 2004

By A Customer

This review is from: Microarrays for an Integrative Genomics (Computational Molecular Biology) (Hardcover)

This is a well written book that gives an overview of the technology of microarrays and their use as investigative tools in functional genomics experiments. I found the technical and analytical descriptions very easy to follow. This is still the only book around that can bring any investigator with little knowledge of molecular biology, data analysis, and/or microarrays up to speed in the field. It is also a good text book for a graduate level course on microarray data analysis.

8 of 10 people found the following review helpful:

4.0 out of 5 stars lots of important stuff, January 20, 2004

By A Customer

This review is from: Microarrays for an Integrative Genomics (Computational Molecular Biology) (Hardcover)

This book contains lots of important topical information on the design and analysis of microarray experiments. It calls attention to a lot of important but sometimes subtle issues that many biologists appear to be overlooking. It appears to be a must-read for researchers who want to avoid expensive dead ends. But it's not perfect...

A well-informed computer scientist will recognize that quite a few computational statements are just plain wrong (e.g., p 180,
"[Dendrograms] require the comprehensive precomputation of the dissimilarity measure for all pairs of genes, which grows on the order of N^2" Wrong! Try bucketing. Or p 139, a dissimilarity function based on linear correlation coefficients is "definite". No! If x is a vector and C is a scalar, then clearly x=/=Cx, but d(x,Cx)=0, contrary to the definition of "definite". The "pseudocode" in Chapter 4 is not any clearer than the text, and it is not structured in a way that would allow it to be elaborated into well-engineered code. So rely on this book for big ideas and references, not for details. The book also reinforces my preconception that MIT Press doesn't employ editors... 'way too many typos, for starters.

You have to know the basics of molecular biology for this book, and it wouldn't hurt to have a basic understanding of DNA chips as well. It's definitely not the first step for a mathematical scientist hoping to become a bioinformatician. (But why should it be? :c)