Automated Annotation of ABC Transporters

The ABC superfamily of transporters is an ancient, diverse and ubiquitous group of transporters, present throughout all kingdoms of life [36 (link)]. This superfamily, which performs both uptake and efflux duties, comprises a large number of families and subfamilies. These are defined based primarily on the significant evolutionary and structural variation in the membrane component, as these are highly conserved, comprising a number of motifs and regions that are well characterized [37–39 (link)].The Pfam family ABC_tran PF00005 (https://www.ebi.ac.uk/interpro/entry/pfam/PF00005/) identifies the highly conserved ATP-binding cassette, which binds and hydrolyses ATP, thereby coupling transport to ATP hydrolysis in a large number of biological processes. Besides transporter ATP-binding cassettes, PF00005 also identifies a group of ATPases, e.g. excinuclease ATPase (UvrA), chromosome segregation ATPase (Smc), recombinational DNA repair ATPase (RecF) and ATPases involved in chromosome partitioning (Soj, Mrp), as well as a group of CBS-domain-containing transcriptional regulators and signal transduction proteins. To remove these false positives, we set up a series of negative rules whereby all putative ABC transporters, based on an inclusive and promiscuous search with PF00005, that also have positive hits to the COGs representing these non-transporter ATPases and CBS-domain-containing proteins (shown in Table 1) are deleted. Furthermore, a text search of the annotation of blastp matches to the NR database is also conducted, and where the top hits are found to contain keywords shown in Table 1, these entries are also removed. These rules assist in the specific identification of ATP-binding cassette transporters.
The submission database is polled every 24 h and new submissions are automatically processed via the TransAAP pipeline on a dedicated Linux cluster, running the relevant searches and predicting the complete transporter contents, then uploading all the results into the TransportDB MySQL database. Users can check the status of the annotation of their submitted genomes and once complete can access a list of predicted transporters, or individual transporter annotation pages. Users can view annotation, supporting evidence, and curate the annotation on each individual transporter annotation page, or download the result as a tab delimited text file (Fig. 1).