With the advent of whole-transcriptome studies and the growing need for public repositories, it has become essential to combine multiple heterogeneous datasets for immune cells. In this chapter, we describe the implementation of a compendium of 10,833 genes for 975 samples, corresponding to 52 resting immune cell types. We begin by describing the datasets, and their selection, in particular. We then explain the methodology implemented to create a qualified compendium: the processing of each array (quality control, normalization and bias correction), integration (merging rules, global normalization and batch removal) and validation. Finally some examples of use will be detailed. The utility and limitations of the compendium are also discussed, as an introduction to the next version.
Keywords: Bioinformatics; Compendium; Gene expression profiling; Immunology; Integration; Transcriptomics.
© 2020 Elsevier Inc. All rights reserved.