ADP-ribosylation is a critical post-translational modification involved in regulating diverse cellular processes, including chromatin structure regulation, RNA transcription, and cell death. It is accomplished by transferring a single or multiple ADP-ribose unit(s) from NAD+ onto target substrates with the release of Nicotinamide via the ADP-ribosyltransferase superfamily. Bacterial ADP-ribosyltransferase toxins (bARTTs) serve as potent virulence factors that orchestrate the manipulation of host cell functions to facilitate bacterial pathogenesis. bARTTs are encoded by various important human pathogens, such as Vibrio cholerae, Bordetella pertussis, Salmonella typhi, Staphylococcus aureus, Pseudomonas aeruginosa, Mycoplasma pneumonia, Corynebacterium diphtheriae and Clostridium botulinum. Despite their pivotal role, the bioinformatic identification of novel bARTTs poses a formidable challenge due to limited verified data and the inherent sequence diversity among bARTT members.

We have proposed a deep learning-based ARTNet specifically engineered to predict bARTTs from bacterial genomes. ARTNet achieved an MCC of 0.9351 and an F1-score (macro) of 0.9666 on repeated independent test datasets, outperforming three other deep learning models and traditional machine learning algorithms in terms of time efficiency and accuracy. Besides, ARTNet has capability to predict novel bARTTs across domain superfamilies without sequence similarity. ARTNet models trained on pos_art_346, pos_art_346_random and pos_whole were applied to build this web server for the prediction of potential bARTTs from protein sequences of interest. This web server provides three modes: comprehensive, medium, and strict, to report positive sequences supported by at least one model, at least two models, and all three models, respectively. Users can submit one or multiple sequences in FASTA format for prediction by single click. The maximum number of sequences in one batch is set to 5000 to avoid abuse and overloading. Users can also download our source code from https://github.com/zhengdd0422/ARTNet/ to perform personalized large-scale sequence predictions.

Zheng D., Zhou S., Chen L., Pang G., Yang J., 2024. A deep learning method to predict bacterial ADP-ribosyltransferase toxins. Bioinformatics 40(7):btae378.

If you have any suggestions or comments to ARTNet, please contact us via: yangj@ipbcams.ac.cn