Code, models, and data needed to reproduce computational figures for the manuscript "Discovery of antibiotics that selectively kill metabolically dormant bacteria" (Zheng, Valeri, Andrews, et al. 2023).
There is a need to discover and develop non-toxic antibiotics that are effective against metabolically dormant bacteria, which underlie chronic infections and promote antibiotic resistance. Traditional antibiotic discovery has historically favored compounds effective against actively metabolizing cells, a property that is not predictive of efficacy in dormant contexts. Here, we combine a stationary-phase screening method with deep learning-powered virtual screens and toxicity filters to evaluate properties of compounds with lethality against dormant bacteria. This work demonstrates the value of harnessing non-traditional screening methods and deep learning models to identify non-toxic antibacterial compounds that are effective in infection-relevant contexts.
All computational analyses can be accessed at this repository, with machine learning models, raw data files, processed output files, source code, and reproduced figures available in full and labelled by figure name.
- Raw data have been deposited in the data/ folder and are publicly available as of the date of submission.
- All original code has been deposited in the src/ folder and is publicly available as of the date of submission.
- Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.
Some raw data files have been zipped so that they can be uploaded in GitHub, or omitted altogether if there was a suitable downstream processed version of the file. Some random forest classifier .pkl files have been omitted due to size reasons and because these models were not ultimately used for any predictive purposes, but feel free to request the RFC .pkl files from us if you need them. If you have any issues, questions, or concerns, please do not hesitate to open up an issue or contact jackievaleri8 "at" gmail "dot" com.