A machine learning model trained on a high-throughput antibacterial screen increases the hit rate of drug discovery.
Screening for novel antibacterial compounds in small molecule libraries has a low success rate. We applied machine learning (ML)-based virtual screening for antibacterial activity and evaluated its predictive power by experimental validation. We first binarized 29,537 compounds according to their gr...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Public Library of Science (PLoS)
2022-10-01
|
| Series: | PLoS Computational Biology |
| Online Access: | https://doi.org/10.1371/journal.pcbi.1010613 |
| _version_ | 1798044707572416512 |
|---|---|
| author | A S M Zisanur Rahman Chengyou Liu Hunter Sturm Andrew M Hogan Rebecca Davis Pingzhao Hu Silvia T Cardona |
| author_facet | A S M Zisanur Rahman Chengyou Liu Hunter Sturm Andrew M Hogan Rebecca Davis Pingzhao Hu Silvia T Cardona |
| author_sort | A S M Zisanur Rahman |
| collection | DOAJ |
| description | Screening for novel antibacterial compounds in small molecule libraries has a low success rate. We applied machine learning (ML)-based virtual screening for antibacterial activity and evaluated its predictive power by experimental validation. We first binarized 29,537 compounds according to their growth inhibitory activity (hit rate 0.87%) against the antibiotic-resistant bacterium Burkholderia cenocepacia and described their molecular features with a directed-message passing neural network (D-MPNN). Then, we used the data to train an ML model that achieved a receiver operating characteristic (ROC) score of 0.823 on the test set. Finally, we predicted antibacterial activity in virtual libraries corresponding to 1,614 compounds from the Food and Drug Administration (FDA)-approved list and 224,205 natural products. Hit rates of 26% and 12%, respectively, were obtained when we tested the top-ranked predicted compounds for growth inhibitory activity against B. cenocepacia, which represents at least a 14-fold increase from the previous hit rate. In addition, more than 51% of the predicted antibacterial natural compounds inhibited ESKAPE pathogens showing that predictions expand beyond the organism-specific dataset to a broad range of bacteria. Overall, the developed ML approach can be used for compound prioritization before screening, increasing the typical hit rate of drug discovery. |
| first_indexed | 2024-04-11T23:08:29Z |
| format | Article |
| id | doaj.art-5a6fd3b2e7f448648e12f1623ece1423 |
| institution | Directory Open Access Journal |
| issn | 1553-734X 1553-7358 |
| language | English |
| last_indexed | 2024-04-11T23:08:29Z |
| publishDate | 2022-10-01 |
| publisher | Public Library of Science (PLoS) |
| record_format | Article |
| series | PLoS Computational Biology |
| spelling | doaj.art-5a6fd3b2e7f448648e12f1623ece14232022-12-22T03:57:55ZengPublic Library of Science (PLoS)PLoS Computational Biology1553-734X1553-73582022-10-011810e101061310.1371/journal.pcbi.1010613A machine learning model trained on a high-throughput antibacterial screen increases the hit rate of drug discovery.A S M Zisanur RahmanChengyou LiuHunter SturmAndrew M HoganRebecca DavisPingzhao HuSilvia T CardonaScreening for novel antibacterial compounds in small molecule libraries has a low success rate. We applied machine learning (ML)-based virtual screening for antibacterial activity and evaluated its predictive power by experimental validation. We first binarized 29,537 compounds according to their growth inhibitory activity (hit rate 0.87%) against the antibiotic-resistant bacterium Burkholderia cenocepacia and described their molecular features with a directed-message passing neural network (D-MPNN). Then, we used the data to train an ML model that achieved a receiver operating characteristic (ROC) score of 0.823 on the test set. Finally, we predicted antibacterial activity in virtual libraries corresponding to 1,614 compounds from the Food and Drug Administration (FDA)-approved list and 224,205 natural products. Hit rates of 26% and 12%, respectively, were obtained when we tested the top-ranked predicted compounds for growth inhibitory activity against B. cenocepacia, which represents at least a 14-fold increase from the previous hit rate. In addition, more than 51% of the predicted antibacterial natural compounds inhibited ESKAPE pathogens showing that predictions expand beyond the organism-specific dataset to a broad range of bacteria. Overall, the developed ML approach can be used for compound prioritization before screening, increasing the typical hit rate of drug discovery.https://doi.org/10.1371/journal.pcbi.1010613 |
| spellingShingle | A S M Zisanur Rahman Chengyou Liu Hunter Sturm Andrew M Hogan Rebecca Davis Pingzhao Hu Silvia T Cardona A machine learning model trained on a high-throughput antibacterial screen increases the hit rate of drug discovery. PLoS Computational Biology |
| title | A machine learning model trained on a high-throughput antibacterial screen increases the hit rate of drug discovery. |
| title_full | A machine learning model trained on a high-throughput antibacterial screen increases the hit rate of drug discovery. |
| title_fullStr | A machine learning model trained on a high-throughput antibacterial screen increases the hit rate of drug discovery. |
| title_full_unstemmed | A machine learning model trained on a high-throughput antibacterial screen increases the hit rate of drug discovery. |
| title_short | A machine learning model trained on a high-throughput antibacterial screen increases the hit rate of drug discovery. |
| title_sort | machine learning model trained on a high throughput antibacterial screen increases the hit rate of drug discovery |
| url | https://doi.org/10.1371/journal.pcbi.1010613 |
| work_keys_str_mv | AT asmzisanurrahman amachinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT chengyouliu amachinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT huntersturm amachinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT andrewmhogan amachinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT rebeccadavis amachinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT pingzhaohu amachinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT silviatcardona amachinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT asmzisanurrahman machinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT chengyouliu machinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT huntersturm machinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT andrewmhogan machinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT rebeccadavis machinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT pingzhaohu machinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery AT silviatcardona machinelearningmodeltrainedonahighthroughputantibacterialscreenincreasesthehitrateofdrugdiscovery |