Computational Text Analysis For Functional Genomics and Bioinformatics

by Raychaudhuri, Soumya

• ISBN13: 9780198567417
• ISBN10: 0198567413
• Edition: 1st
• Format: Paperback
• Copyright: 2006-03-30
• Publisher: Oxford University Press

Summary

This book brings together the two disparate worlds of computational text analysis and biology and presents some of the latest methods and applications to proteomics, sequence analysis and gene expression data. Modern genomics generates large and comprehensive data sets but their interpretation requires an understanding of a vast number of genes, their complex functions, and interactions. Keeping up with the literature on a single gene is a challenge itself-for thousands of genes it is simply impossible. Here, Soumya Raychaudhuri presents the techniques and algorithms needed to access and utilize the vast scientific text, i.e. methods that automatically "read" the literature on all the genes. Including background chapters on the necessary biology, statistics and genomics, in addition to practical examples of interpreting many different types of modern experiments, this book is ideal for students and researchers in computational biology, bioinformatics, genomics, statistics and computer science.

Table of Contents

List of Figures	xvii
List of Plates	xxi
List of Tables	xxiii
An introduction to text analysis in genomics	1	(16)
The genomics literature	2	(3)
Using text in genomics	5	(4)
Building databases of genetic knowledge	5	(2)
Analyzing experimental genomic data sets	7	(1)
Proposing new biological knowledge: identifying candidate genes	8	(1)
Publicly available text resources	9	(3)
Electronic text	9	(1)
Genome resources	9	(2)
Gene ontology	11	(1)
The advantage of text-based methods	12	(1)
Guide to this book	13	(4)
Functional genomics	17	(66)
Some molecular biology	17	(10)
Central dogma of molecular biology	18	(1)
Deoxyribonucleic acid	18	(2)
Ribonucleic acid	20	(2)
Genes	22	(2)
Proteins	24	(2)
Biological function	26	(1)
Probability theory and statistics	27	(10)
Probability	27	(1)
Conditional probability	28	(1)
Independence	29	(1)
Bayes' theorem	30	(1)
Probability distribution functions	31	(2)
Information theory	33	(1)
Population statistics	34	(1)
Measuring performance	35	(2)
Deriving and analyzing sequences	37	(24)
Sequencing	39	(1)
Homology	40	(2)
Sequence alignment	42	(2)
Pairwise sequence alignment and dynamic programming	44	(3)
Linear time pairwise alignment: BLAST	47	(1)
Multiple sequence alignment	48	(2)
Comparing sequences to profiles: weight matrices	50	(3)
Position specific iterative BLAST	53	(1)
Hidden Markov models	54	(7)
Gene expression profiling	61	(22)
Measuring gene expression with arrays	63	(1)
Measuring gene expression by sequencing and counting transcripts	64	(1)
Expression array analysis	65	(1)
Unsupervised grouping: clustering	66	(2)
K-means clustering	68	(1)
Self-organizing maps	69	(1)
Hierarchical clustering	70	(2)
Dimension reduction with principal components analysis	72	(2)
Combining expression data with external information: supervised machine learning	74	(1)
Nearest neighbor classification	75	(1)
Linear discriminant analysis	75	(8)
Textual profiles of genes	83	(24)
Representing documents as word vectors	84	(2)
Metrics to compare documents	86	(2)
Some words are more important for document similarity	88	(1)
Building a vocabulary: feature selection	88	(2)
Weighting words	90	(2)
Latent semantic indexing	92	(2)
Defining textual profiles for genes	94	(2)
Using text like genomics data	96	(4)
A simple strategy to assigning keywords to groups of genes	100	(1)
Querying genes for biological function	101	(6)
Using text in sequence analysis	107	(16)
SWISS-PROT records as a textual resource	109	(2)
Using sequence similarity to extend literature references	111	(1)
Assigning keywords to summarize sequences hits	112	(2)
Using textual profiles to organize sequence hits	114	(1)
Using text to help identify remote homology	114	(1)
Modifying iterative sequence similarity searches to include text	115	(2)
Evaluating PSI-BLAST modified to include text	117	(3)
Combining sequence and text together	120	(3)
Text-based analysis of a single series of gene expression measurements	123	(24)
Pitfalls of gene expression analysis: noise	124	(2)
Phosphate metabolism: an example	126	(1)
The top fifteen genes	127	(2)
Distinguishing true positives from false positives with a literature-based approach	129	(1)
Neighbor expression information	130	(2)
Application to phosphate metabolism data set	132	(4)
Recognizing high induction false positives with literature-based scores	136	(2)
Recognizing low induction false positives	138	(2)
Assessing experiment quality with literature-based scoring	140	(1)
Improvements	140	(1)
Application to other assays	141	(1)
Assigning keywords that describe the broad biology of the experiment	141	(6)
Analyzing groups of genes	147	(24)
Functional coherence of a group of genes	148	(4)
Overview of computational approach	152	(3)
Strategy to evaluate different algorithms	155	(2)
Word distribution divergence	157	(3)
Best article score	160	(3)
Neighbor divergence	163	(1)
Calculating a theoretical distribution of scores	163	(1)
Quantifying the difference between the empirical score distribution and the theoretical one	164	(1)
Neighbor divergence per gene	164	(2)
Corruption studies	166	(1)
Application of functional coherence scoring to screen gene expression clusters	167	(3)
Understanding the gene group's function	170	(1)
Analyzing large gene expression data sets	171	(24)
Groups of genes	172	(1)
Assigning keywords	173	(1)
Screening gene expression clusters	173	(5)
Optimizing cluster boundaries: hierarchical clustering	178	(6)
Application to other organisms besides yeast	184	(5)
Identifying and optimizing clusters in a Drosophila development data set	189	(6)
Using text classification for gene function annotation	195	(32)
Functional vocabularies and gene annotation	196	(6)
Gene Ontology	197	(3)
Enzyme Commission	200	(1)
Kyoto Encyclopedia of Genes and Genomes	200	(2)
Text classification	202	(1)
Nearest neighbor classification	203	(1)
Naive Bayes classification	204	(1)
Maximum entropy classification	205	(5)
Feature selection: choosing the best words for classification	210	(2)
Classifying documents into functional categories	212	(1)
Comparing classifiers	213	(8)
Annotating genes	221	(6)
Finding gene names	227	(18)
Strategies to identify gene names	228	(1)
Recognizing gene names with a dictionary	228	(4)
Using word structure and appearance to identify gene names	232	(1)
Using syntax to eliminate gene name candidates	233	(2)
Using context as a clue about gene names	235	(2)
Morphology	237	(1)
Identifying gene names and their abbreviations	237	(3)
A single unified gene name finding algorithm	240	(5)
Protein interaction networks	245	(26)
Genetic networks	246	(1)
Experimental assays to identify protein networks	247	(2)
Yeast two hybrid	247	(1)
Affinity precipitation	248	(1)
Predicting interactions versus verifying interactions with scientific text	249	(1)
Networks of co-occurring genes	249	(1)
Protein interactions and gene name co-occurrence in text	250	(4)
Number of textual co-occurrences predicts likelihood of an experimentally predicted interaction	254	(5)
Information extraction and genetic networks: increasing specificity and identifying interaction type	259	(3)
Statistical machine learning	262	(9)
Conclusion	271	(2)
Index	273