Shira L. Broschat
Shirley W. I. Siu
Cesar de la Fuente-Nunez- 1School of Electrical Engineering and Computer Science, Washington State University, Pullman, WA, United States
- 2Department of Veterinary Microbiology and Pathology, Washington State University, Pullman, WA, United States
- 3Paul G. Allen School for Global Health, Washington State University, Pullman, WA, United States
- 4Centre for Artificial Intelligence Driven Drug Discovery, Macao Polytechnic University, Macao, Macao SAR, China
- 5Faculty of Applied Sciences, Macao Polytechnic University, Macao, Macao SAR, China
- 6Machine Biology Group, Departments of Psychiatry and Microbiology, Institute for Biomedical Informatics, Institute for Translational Medicine and Therapeutics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, United States
- 7Department of Bioengineering, University of Pennsylvania, Philadelphia, PA, United States
- 8Department of Chemical and Biomolecular Engineering, University of Pennsylvania, Philadelphia, PA, United States
- 9Department of Chemistry, University of Pennsylvania, Philadelphia, PA, United States
- 10Penn Institute for Computational Science, University of Pennsylvania, Philadelphia, PA, United States
Editorial on the Research Topic
Machine learning approaches to antimicrobials: discovery and resistance
In 1900, infectious diseases were responsible for nearly 60% of deaths in the United States. By 2010, this had dropped to only 1.62% (Jones et al., 2012). The reason for this dramatic decrease was the introduction of antibiotics in the mid-20th century. However, antimicrobial resistance (AMR) now jeopardizes this achievement. In fact, AMR is a major public health threat with millions of AMR infections occurring annually, often involving resistance to multiple antibiotics. In the United States, AMR accounts for more than 35,000 deaths each year and over 1.27 million deaths globally (CDC, 2024). Unfortunately, because of the high cost of development and lack of return on investment, large pharmaceutical companies are unwilling to commit to developing new antimicrobials to combat AMR without incentives (Dutescu and Hillier, 2021; WHO, 2023). Reducing the expense and uncertainty associated with conventional methods of drug discovery may serve as a powerful incentive.
Machine learning (ML), a branch of artificial intelligence, can assist in the fight against AMR with the development of new antimicrobials (Torres et al., 2022; Maasch et al., 2023; Wong et al., 2023; Santos-Júnior et al., 2024; Wan et al., 2024). It has been used to facilitate drug discovery for many years (Vamathevan et al., 2019). Antimicrobial peptides (AMPs) can be developed as antibiotics, but conventional methods of discovery are not cost-effective. ML algorithms can be developed to predict AMPs, narrowing the number of possible candidates, which can then be tested experimentally for their efficacy (Wan et al., 2024). ML algorithms can also be used as an alternative approach to sequence similarity to predict AMR genes when they have little similarity to known AMR genes (Chowdhury et al., 2019a; Chowdhury et al., 2019b; Chowdhury et al., 2020). These predictions may then be used to help in the treatment of patients.
This Research Topic, “Machine Learning Approaches to Antimicrobials: Discovery and Resistance,” consists of five papers: three original research papers and two technology and code papers. One paper applies ML to drug discovery while the others address AMR, as discussed below.
Bacteriocins are antimicrobial peptides produced by bacteria to kill other bacteria. Interest in them has grown because of their ability to control AMR bacteria and their possible use as both broad- and narrow-spectrum antibiotics (Sugrue et al., 2024). In their paper, Akhter and Miller introduce BPAGS, a web application for use in predicting bacteriocins. They consider several different feature evaluation methods and show that the best results are obtained with features reduced using an alternating decision tree algorithm with a support vector machine model. They achieve an accuracy of 99.11% and an AUC of 0.9984 for their test data.
Antibiotic resistance genes (ARGs) are found in bacteria, but in addition to those that occur naturally or are obtained via horizontal gene transfer, point mutations can occur in the bacterial chromosome which can give rise to AMR. These variants of ARGs (ARGVs) can be difficult to identify. Marini et al. propose an approach for identifying ARGVs called KARGVA, for k-mer antibiotic resistance gene variant analyzer, which is an extension of their earlier approach KARGA used to identify ARGs. KARGVA performs well on semi-synthetic data, achieving an accuracy of 99.2% for a base rate change of 1%. KARGVA is open-source and publicly available.
Helicobacter pylori is a bacterium that infects the stomach and can cause gastritis, ulcers, and even stomach cancer. Unfortunately, H. pylori is becoming resistant to several antibiotics (Boyanova et al., 2023). In their study, Yu et al. perform whole-genome sequencing (WGS) for 52 strains of H. pylori obtained from 52 different patients to determine their AMR genotypes. The susceptibility of these strains to five different antibiotics is determined experimentally. The authors then use different ML models, including a support vector machine model, with feature selection to predict AMR for the different genotypes and achieve 100% specificity for resistance to amoxicillin and clarithromycin.
Gao et al. also explore the use of WGS in their exploration of the multi-drug resistant bacterium Acinetobacter baumannii, which causes many nosocomial infections. While A. baumannii has low virulence, those who are immunocompromised or elderly and infirm are at risk, especially if they are ICU patients or have lengthy hospital stays (Ayoub Moubareck and Hammoudi Halat, 2020). Gao et al. combine k-mer features from the WGS data with three different ML algorithms to predict AMR to thirteen different drugs. They achieve an average accuracy of 96% for their test data when they use the top 11-mers.
In the final paper on AMR prediction, López-Cortés et al. turn to deep neural networks (DNNs), with data on bacteria obtained via mass spectrometry, to predict AMR to various antibiotics for three different pathogens: Escherichia coli, Staphylococcus aureus, and Klebsiella pneumoniae. Their DNN model, MSDeepAMR, improves upon previous results obtained, and as they point out, their results for predicting resistance to ciprofloxacin are good for all three bacteria.
The articles presented in this Research Topic demonstrate how ML models can be used with a diversity of data to predict antimicrobial peptides in the form of bacteriocins or to predict AMR to numerous antibiotics for a variety of bacteria. They illustrate how models based on ML can be used to assist in combatting AMR.
Author contributions
SB: Writing–original draft, Writing–review and editing. SS: Writing–review and editing. CF-N: Writing–review and editing.
Funding
The author(s) declare financial support was received for the research, authorship, and/or publication of this article. CF-N holds a Presidential Professorship at the University of Pennsylvania, is a recipient of the Langer Prize by the AIChE Foundation, and acknowledges funding from the IADR Innovation in Oral Care Award, the Procter and Gamble Company, United Therapeutics, a BBRF Young Investigator Grant, the Nemirovsky Prize, Penn Health-Tech Accelerator Award, the Dean–s Innovation Fund from the Perelman School of Medicine at the University of Pennsylvania, the National Institute of General Medical Sciences of the National Institutes of Health under award number R35GM138201, and the Defense Threat Reduction Agency (DTRA; HDTRA1-22-10031, HDTRA1-21-1-0014, and HDTRA1-23-1-0001).
Conflict of interest
CF-N provides consulting services to Invaio Sciences and is a member of the Scientific Advisory Boards of Nowture S. L., Peptidus, and Phare Bio. CF-N is also on the Advisory Board of the Peptide Drug Hunting Consortium. The de la Fuente Lab has received research funding or in-kind donations from United Therapeutics, Strata Manufacturing PJSC, and Procter & Gamble, none of which were used in support of this work.
The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
Ayoub Moubareck, C., and Hammoudi Halat, D. (2020). Insights into Acinetobacter baumannii: a review of microbiological, virulence, and resistance traits in a threatening nosocomial pathogen. Antibiotics 9, 119. doi:10.3390/antibiotics9030119
Boyanova, L., Hadzhiyski, P., Gergova, R., and Markovska, R. (2023). Evolution of Helicobacter pylori resistance to antibiotics: a topic of increasing concern. Antibiotics 12, 332. doi:10.3390/antibiotics12020332
CDC (2024). Antimicrobial resistance facts and stats. Available at: https://www.cdc.gov/antimicrobial-resistance/data-research/facts-stats/.
Chowdhury, A. S., Call, D. R., and Broschat, S. L. (2019a). Antimicrobial resistance prediction for gram-negative bacteria via game theory-based feature evaluation. Sci. Rep. 9, 14487. doi:10.1038/s41598-019-50686-z
Chowdhury, A. S., Call, D. R., and Broschat, S. L. (2020). PARGT: a software tool for predicting antimicrobial resistance in bacteria. Sci. Rep. 10, 11033. doi:10.1038/s41598-020-67949-9
Chowdhury, A. S., Khaledian, E., and Broschat, S. L. (2019b). Capreomycin resistance prediction in two species of Mycobacterium using a stacked ensemble method. J. Appl. Micro. 127, 1656–1664. doi:10.1111/jam.14413
Dutescu, I. A., and Hillier, S. A. (2021). Encouraging the development of new antibiotics: are financial incentives the right way forward? A systematic review and case study. Infect. Drug Resist. 14, 415–434. doi:10.2147/IDR.S287792
Jones, D. S., Podolsky, S. H., and Greene, J. A. (2012). The burden of disease and the changing task of medicine. N. Engl. J. Med. 366, 2333–2338. doi:10.1056/NEJMp1113569
Maasch, J. R. M. A., Torres, M. D. T., Melo, M. C. R., and de la Fuente-Nunez, C. (2023). Molecular de-extinction of ancient antimicrobial peptides enabled by machine learning. Cell Host Microbe 31, 1260–1274.e6. doi:10.1016/j.chom.2023.07.001
Santos-Júnior, C. D., Torres, M. D. T., Duan, Y., Rodríguez del Río, Á., Schmidt, T. S., Chong, H., et al. (2024). Discovery of antimicrobial peptides in the global microbiome with machine learning. Cell 187, 3761–3778.e16. doi:10.1016/j.cell.2024.05.013
Sugrue, I., Ross, R. P., and Hill, C. (2024). Bacteriocin diversity, function, discovery and application as antimicrobials. Nat. Rev. Microbiol. doi:10.1038/s41579-024-01045-x
Torres, M. D. T., Melo, M. C. R., Flowers, L., Crescenzi, O., Notomista, E., and de la Fuente-Nunez, C. (2022). Mining for encrypted peptide antibiotics in the human proteome. Nat. Biomed. Eng. 6 (1), 67–75. doi:10.1038/s41551-021-00801-1
Vamathevan, J., Clark, D., Czodrowski, P., Dunham, I., Ferran, E., Lee, G., et al. (2019). Applications of machine learning in drug discovery and development. Nat. Rev. Drug Discov. 18, 463–477. doi:10.1038/s41573-019-0024-5
Wan, F., Torres, M. D. T., Peng, J., and de la Fuente-Nunez, C. (2024). Deep-learning-enabled antibiotic discovery through molecular de-extinction. Nat. Biomed. Eng. doi:10.1038/s41551-024-01201-x
Wan, F., Wong, F., Collins, J. J., et al. (2024). Machine learning for antimicrobial peptide identification and design. Nat. Rev. Bioeng. 2, 392–407. doi:10.1038/s44222-024-00152-x
WHO (2023). Executive summary of “incentivising the development of new antibacterial treatments: progress report by the global AMR R&D hub and WHO 2023”. Available at: https://cdn.who.int/media/docs/default-source/2021-dha-docs/incentivising-development-of-new-antibacterial-treatments-2023---exec-summary.pdf.
Keywords: machine learning, deep learning, antimicrobial resistance, antimicrobial peptides, bacteriocins, WGS, infectious diseases
Citation: Broschat SL, Siu SWI and de la Fuente-Nunez C (2024) Editorial: Machine learning approaches to antimicrobials: discovery and resistance. Front. Bioinform. 4:1458237. doi: 10.3389/fbinf.2024.1458237
Received: 02 July 2024; Accepted: 22 July 2024;
Published: 09 August 2024.
Edited and reviewed by:
Garrett M. Morris, University of Oxford, United KingdomCopyright © 2024 Broschat, Siu and de la Fuente-Nunez. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Shira L. Broschat, shira@wsu.edu