AbstractCharacterization of gene essentiality across different conditions is a useful approach for predicting gene function. Transposon sequencing (TnSeq) is a powerful means of generating genome-wide profiles of essentiality and has been used extensively in Mycobacterium tuberculosis (Mtb) genetic research. Over the past two decades, dozens of TnSeq screens have been published, yielding valuable insights into the biology of Mtb in vitro, inside macrophages, and in model host organisms. However, these Mtb TnSeq profiles are distributed across dozens of research papers within supplementary materials, which makes querying them cumbersome and assembling a complete and consistent synthesis of existing data challenging. Here, we address this problem by building a central repository of publicly available TnSeq screens performed in M. tuberculosis, which we call the Mtb transposon sequencing database (MtbTnDB). The MtbTnDB encompasses 64 published and unpublished TnSeq screens, and is standardized, open-access, and allows users easy access to data, visualizations, and functional predictions through an interactive web-app (www.mtbtndb.app). We also present evidence that (i) genes in the same genomic neighborhood tend to have similar TnSeq profiles, and (ii) clusters of genes with similar TnSeq profiles tend to be enriched for genes belonging to the same functional categories. Finally, we test and evaluate machine learning models trained on TnSeq profiles to guide functional annotation of orphan genes in Mtb. In addition to facilitating the exploration of conditional genetic essentiality in this important human pathogen via a centralized TnSeq data repository, the MtbTnDB will enable hypothesis generation and the extraction of meaningful patterns by facilitating the comparison of datasets across conditions. This will provide a basis for insights into the functional organization of Mtb genes as well as gene function prediction.
A large-scale zebrafish gene knockout resource for the genome-wide study of gene function
