The BSGatlas: An enhanced annotation of genes and transcripts for the Bacillus subtilis genome with improved information access

Abstract

The genome of Bacillus subtilis continues to provide exiting genomic insights. However, the growing collective genomic knowledge about this micro-organism is spread across multiple annotation resources. Thus, the full annotation is not directly accessible neither for specific genes nor for large-scale high-throughput analyses. Furthermore, access to annotation of non-coding RNA genes (ncRNAs) and polycistronic mRNAs is difficult. To address these challenges we introduce the Bacillus subtilis genome atlas, BSGatlas, in which we integrate and unify multiple existing annotation resources. Our integration provides twice as many ncRNAs than the individual resources, improves the positional annotation for 70% of the combined ncRNAs, and makes it possible to infer specific ncRNA types. Moreover, we unify known transcription start sites, termination, and transcriptional units (TUs) as a comprehensive transcript map. This transcript map implies 815 new TUs and 6, 164 untranslated regions (UTRs), which is a five-fold increase over existing resources. We furthermore, find 2, 309 operons covering the transcriptional annotation for 93% of all genes, corresponding to an improvement by 11%. The BSGatlas is available in multiple formats. A user can either download the entire annotation in the standardized GFF3 format, which is compatible with most bioinformatics tools for omics and high-throughput studies, or view the annotation in an online browser at http://rth.dk/resources/bsgatlas.\n\nImportanceThe Bacillus subtilis genome has been studied in numerous context and consequently multiple efforts have been made in providing a complete annotation. Unfortunately, a number of resources are no longer maintained, and (i) the collective annotation knowledge is dispersed over multiple resources, of which each has a different focus of what type of annotation information they provide. (ii) Thus, it is difficult to easily and at a large scale obtain information for a genomic region or genes of interest. (iii) Furthermore, all resources are essentially incomplete when it comes to annotating non-coding and structured RNA, and transcripts in general. Here, we address all three problems by first collecting existing annotations of genes and transcripts start and termination sites; afterwards resolving discrepancies in annotations and combining them, which doubled the number of ncRNAs; inferring full transcripts and 2,309 operons from the combined knowledge of known transcript boundaries and meta-information; and critically providing it all in a standardized UCSC browser. That interface and its powerful set of functionalities allow users to access all the information in a single resource as well as enables them to include own data on top the full annotation.