Large-scale transcriptomic studies have been instrumental in characterising gene expression in the human brain across brain regions, developmental stages and disease states. Several large-scale consortia have generated gene expression data from the human brain including GTEx, PsychENCODE, CommonMind, BrainSeq and BrainSpan. The data generated by these consortia partially overlaps in terms of sample characteristics, and has been generated with different library preparation and data analysis methods.
Despite this wealth of data, it is currently difficult to extract and compare gene expression information across these large-scale datasets. This difficulty is further exemplified by inconsistent and incomplete annotation of metadata for a given dataset. Generally complete annotations include all aspects of experimental data (e.g RNA quality, library preparation) and computational pipelines (e.g software arguments and versions) to allow the researcher to correlate these variables with gene expression. However, lack of such information may compromise the ability to reproduce the results of the original study.
Here, we present the Brain Integrative Transcriptome Hub (BITHub), a web resource that aggregates gene-expression data from the human brain across multiple consortia, and allows direct comparison of gene expression in an interactive manner. BITHub contains gene expression data from several large-scale bulk and single-nucleus RNA-seq (snRNA-seq) datasets, with the bulk data totalling to 5,550 samples across 31 regions, and snRNA-seq data to 128,419 cells across 9 major cell-types in the human brain.
BITHub also allows the exploration of covariates that drive gene expression variance in each dataset. We believe that this user-friendly web resource will be valuable for neuroscience and neurogenetics research by allowing biologists and clinical geneticists to uncover patterns of gene expression in the human brain, and to determine whether observations are replicable across datasets.