Genomics Data Analysis: False Discovery Rates and Empirical Bayes Methods

Statisticians have developed empirical Bayes methods to test hundreds or thousands of genomics hypotheses simultaneously, combining advantages of traditional Bayesian and frequentist statistics. This book provides an elementary introduction to these methods, with examples focusing on gene expression, genetic association, and metabolomics and proteomics data. It covers choosing between multiple hypothesis testing methods, converting software output to local false discovery rates, and addressing criticisms of p-values.

Format: Paperback / softback
Length: 140 pages
Publication date: 21 January 2023
Publisher: Taylor & Francis Ltd

Statisticians have revolutionized the field of genomics by developing novel empirical Bayes methods that combine the strengths of traditional Bayesian and frequentist statistics. These methods enable the simultaneous testing of hundreds or thousands of genomics hypotheses, providing a powerful tool for uncovering biological insights.

One of the key innovations of these methods is the estimation of the local false discovery rate (FDR). FDR is a measure of the probability of falsely rejecting a true null hypothesis, and it plays a crucial role in determining the significance of genetic associations and differential gene expression. Traditional FDR methods rely on subjective prior distributions, which can be challenging to interpret and may introduce bias.

However, the new empirical Bayes methods overcome this limitation by assigning probabilities of differential gene expression, genetic association, and other outcomes without requiring subjective prior distributions. This approach allows for more objective and reliable estimates of FDR, which can be used to assess the strength of evidence against a null hypothesis.

This book is designed to bring these innovative methods to scientists and researchers, while keeping the mathematics at an elementary level. It provides a comprehensive introduction to local FDR, covering topics such as the fundamental concepts, dice games and exercises, examples focusing on gene expression, genetic association, metabolomics, and proteomics data, and a gradual introduction to the mathematical equations needed. The book also offers guidance on choosing between different methods of multiple hypothesis testing, converting the output of genomics hypothesis testing software to estimates of local FDR, and navigating through the complexities of current criticisms of p-values.

Furthermore, the book delves into non-Bayesian prior p-values and posterior p-values, providing valuable insights into alternative approaches to statistical inference. This material is not commonly published in other resources, making it a valuable resource for researchers seeking to expand their knowledge in genomics and statistical analysis.

In conclusion, this book is a must-read for scientists and researchers interested in genomics and statistical analysis. It provides a comprehensive and accessible introduction to local FDR, enabling researchers to analyze their own genomics data and critically evaluate published genomics research. With its innovative methods, practical examples, and clear explanations, this book will empower scientists to make informed decisions based on rigorous statistical analysis.

Weight: 260g
Dimension: 216 x 138 (mm)
ISBN-13: 9781032475288