What This Document Is
This document presents detailed instructional content for Statistical Genetics (STATISTICS 246) at the University of California, Berkeley, specifically focusing on the challenges of multiple testing within the context of large-scale gene expression experiments. It delves into the statistical considerations necessary when analyzing data generated from technologies used to measure the expression levels of thousands of genes simultaneously. The material explores methods for interpreting results and drawing valid conclusions when conducting numerous statistical tests.
Why This Document Matters
This resource is invaluable for students enrolled in advanced biostatistics or statistical genetics courses. It’s particularly helpful for those preparing to analyze high-throughput genomic data, such as microarray or RNA-seq data. Researchers involved in gene expression studies, differential gene expression analysis, and genomic data interpretation will also find this material beneficial. Understanding these concepts is crucial for avoiding common pitfalls in statistical inference and ensuring the reliability of research findings. Accessing the full content will equip you with the tools to confidently navigate complex genomic datasets.
Topics Covered
* The motivation for multiple testing correction in gene expression analysis.
* Univariate hypothesis testing as a foundation for broader analysis.
* Methods for adjusting p-values to account for multiple comparisons.
* Concepts related to Type I and Type II errors in the context of genomic data.
* Different error rates used in multiple testing (PCER, PFER, FWER, FDR, pFDR).
* The distinction between strong and weak control in multiple testing procedures.
What This Document Provides
* A discussion of the statistical goals when analyzing gene expression data.
* Real-world examples illustrating the need for multiple testing correction, including studies on Apo Al knock-out mice and leukemia.
* A framework for understanding the challenges of interpreting p-values when testing thousands of genes.
* An exploration of permutation-based methods for estimating p-values.
* A conceptual overview of various error rate control strategies.
* A foundation for further study in advanced statistical genetics techniques.
