Brain Data Alchemy Project: Meta-Analysis of Re-Analyzed Public Transcriptional Profiling Data in the Gemma Database v1

Abstract

Over the past two decades, transcriptional profiling has become an increasingly common tool for investigating the effects of diseases and experimental manipulations on the nervous system. Within transcriptional profiling experiments, microarray or sequencing technologies are used to measure the relative amount of RNA transcript for each of the thousands of genes expressed in a sample. The primary objective of these experiments is to identify genes that are differentially expressed in response to conditions of interest. However, transcriptional profiling experiments have traditionally been conducted with small sample sizes due to expense (e.g., n=3-10/group), resulting in low statistical power. Due to low power, these experiments are prone to capturing large technical artifacts and false positives rather than smaller biological effects of interest. To address this issue, we developed a 10-week summer research program (The Brain Data Alchemy Project) that guides participants through the process of performing a meta-analysis of differential expression effect sizes (Log2 Fold Change or Log2FC) extracted from publicly available transcriptional profiling datasets. To conduct our meta-analyses, we leverage the efforts of the Gemma project, which has curated, preprocessed, and re-analyzed over 19,000 publicly available datasets (https://gemma.msl.ubc.ca/home.html). Participants learn the fundamental principles of systematic review and R programming to conduct the dataset search, result extraction, and whole transcriptome meta-analysis. This protocol outlines the methods used during the first pilot year of the program in 2022.