Reinforcement learning informs optimal treatment strategies to limit antibiotic resistance.

Antimicrobial resistance was estimated to be associated with 4.95 million deaths worldwide in 2019. It is possible to frame the antimicrobial resistance problem as a feedback-control problem. If we could optimize this feedback-control problem and translate our findings to the clinic, we could slow, prevent, or reverse the development of high-level drug resistance. Prior work on this topic has relied on systems where the exact dynamics and parameters were known a priori. In this study, we extend this work using a reinforcement learning (RL) approach capable of learning effective drug cycling policies in a system defined by empirically measured fitness landscapes. Crucially, we show that it is possible to learn effective drug cycling policies despite the problems of noisy, limited, or delayed measurement. Given access to a panel of 15 [Formula: see text]-lactam antibiotics with which to treat the simulated Escherichia coli population, we demonstrate that RL agents outperform two naive treatment paradigms at minimizing the population fitness over time. We also show that RL agents approach the performance of the optimal drug cycling policy. Even when stochastic noise is introduced to the measurements of population fitness, we show that RL agents are capable of maintaining evolving populations at lower growth rates compared to controls. We further tested our approach in arbitrary fitness landscapes of up to 1,024 genotypes. We show that minimization of population fitness using drug cycles is not limited by increasing genome size. Our work represents a proof-of-concept for using AI to control complex evolutionary processes.

A low-footprint, fluorescence-based bacterial time-kill assay for estimating dose-dependent cell death dynamics.

Dose-response curves that describe the relationship between antibiotic dose and growth rate in bacteria are commonly measured with optical density (OD) based assays. While being simple and high-throughput, any dose-dependent cell death dynamics are obscured, as OD assays in batch culture can only quantify a positive net change in cells. Time-kill experiments can be used to quantify cell death rates, but current techniques are extremely resource-intensive and may be biased by residual drug carried over into the quantification assay. Here, we report a novel, fluorescence-based time-kill assay leveraging resazurin as a viable cell count indicator. Our method improves upon previous techniques by greatly reducing the material cost and being robust to residual drug carry-over. We demonstrate our technique by quantifying a dose-response curve in Escherichia coli subject to cefotaxime, revealing dose-dependent death rates. We also show that our method is robust to extracellular debris and cell aggregation. Dose-response curves quantified with our method may provide a more accurate description of pathogen response to therapy, paving the way for more accurate integrated pharmacodynamic-pharmacokinetic studies.

Diverse mutant selection windows shape spatial heterogeneity in evolving populations.

Mutant selection windows (MSWs), the range of drug concentrations that select for drug-resistant mutants, have long been used as a model for predicting drug resistance and designing optimal dosing strategies in infectious disease. The canonical MSW model offers comparisons between two subtypes at a time: drug-sensitive and drug-resistant. In contrast, the fitness landscape model with N alleles, which maps genotype to fitness, allows comparisons between N genotypes simultaneously, but does not encode continuous drug response data. In clinical settings, there may be a wide range of drug concentrations selecting for a variety of genotypes in both cancer and infectious diseases. Therefore, there is a need for a more robust model of the pathogen response to therapy to predict resistance and design new therapeutic approaches. Fitness seascapes, which model genotype-by-environment interactions, permit multiple MSW comparisons simultaneously by encoding genotype-specific dose-response data. By comparing dose-response curves, one can visualize the range of drug concentrations where one genotype is selected over another. In this work, we show how N-allele fitness seascapes allow for N * 2N-1 unique MSW comparisons. In spatial drug diffusion models, we demonstrate how fitness seascapes reveal spatially heterogeneous MSWs, extending the MSW model to more fully reflect the selection of drug resistant genotypes. Furthermore, using synthetic data and empirical dose-response data in cancer, we find that the spatial structure of MSWs shapes the evolution of drug resistance in an agent-based model. By simulating a tumor treated with cyclic drug therapy, we find that mutant selection windows introduced by drug diffusion promote the proliferation of drug resistant cells. Our work highlights the importance and utility of considering dose-dependent fitness seascapes in evolutionary medicine.

Diverse mutant selection windows shape spatial heterogeneity in evolving populations.

Mutant selection windows (MSWs), the range of drug concentrations that select for drug-resistant mutants, have long been used as a model for predicting drug resistance and designing optimal dosing strategies in infectious disease. The canonical MSW model offers comparisons between two subtypes at a time: drug-sensitive and drug-resistant. In contrast, the fitness landscape model with N alleles, which maps genotype to fitness, allows comparisons between N genotypes simultaneously, but does not encode continuous drug response data. In clinical settings, there may be a wide range of drug concentrations selecting for a variety of genotypes. Therefore, there is a need for a more robust model of the pathogen response to therapy to predict resistance and design new therapeutic approaches. Fitness seascapes, which model genotype-by-environment interactions, permit multiple MSW comparisons simultaneously by encoding genotype-specific dose-response data. By comparing dose-response curves, one can visualize the range of drug concentrations where one genotype is selected over another. In this work, we show how N-allele fitness seascapes allow for N*2N-1 unique MSW comparisons. In spatial drug diffusion models, we demonstrate how fitness seascapes reveal spatially heterogeneous MSWs, extending the MSW model to more accurately reflect the selection fo drug resistant genotypes. Furthermore, we find that the spatial structure of MSWs shapes the evolution of drug resistance in an agent-based model. Our work highlights the importance and utility of considering dose-dependent fitness seascapes in evolutionary medicine.

Reinforcement Learning informs optimal treatment strategies to limit antibiotic resistance.

Antimicrobial resistance was estimated to be associated with 4.95 million deaths worldwide in 2019. It is possible to frame the antimicrobial resistance problem as a feedback-control problem. If we could optimize this feedback-control problem and translate our findings to the clinic, we could slow, prevent or reverse the development of high-level drug resistance. Prior work on this topic has relied on systems where the exact dynamics and parameters were known a priori. In this study, we extend this work using a reinforcement learning (RL) approach capable of learning effective drug cycling policies in a system defined by empirically measured fitness landscapes. Crucially, we show that is possible to learn effective drug cycling policies despite the problems of noisy, limited, or delayed measurement. Given access to a panel of 15 ß-lactam antibiotics with which to treat the simulated E. coli population, we demonstrate that RL agents outperform two naive treatment paradigms at minimizing the population fitness over time. We also show that RL agents approach the performance of the optimal drug cycling policy. Even when stochastic noise is introduced to the measurements of population fitness, we show that RL agents are capable of maintaining evolving populations at lower growth rates compared to controls. We further tested our approach in arbitrary fitness landscapes of up to 1024 genotypes. We show that minimization of population fitness using drug cycles is not limited by increasing genome size. Our work represents a proof-of-concept for using AI to control complex evolutionary processes.

UV decontamination of personal protective equipment with idle laboratory biosafety cabinets during the COVID-19 pandemic.

Personal protective equipment (PPE) is crucially important to the safety of both patients and medical personnel, particularly in the event of an infectious pandemic. As the incidence of Coronavirus Disease 2019 (COVID-19) increases exponentially in the United States and many parts of the world, healthcare provider demand for these necessities is currently outpacing supply. In the midst of the current pandemic, there has been a concerted effort to identify viable ways to conserve PPE, including decontamination after use. In this study, we outline a procedure by which PPE may be decontaminated using ultraviolet (UV) radiation in biosafety cabinets (BSCs), a common element of many academic, public health, and hospital laboratories. According to the literature, effective decontamination of N95 respirator masks or surgical masks requires UV-C doses of greater than 1 Jcm-2, which was achieved after 4.3 hours per side when placing the N95 at the bottom of the BSCs tested in this study. We then demonstrated complete inactivation of the human coronavirus NL63 on N95 mask material after 15 minutes of UV-C exposure at 61 cm (232 µWcm-2). Our results provide support to healthcare organizations looking for methods to extend their reserves of PPE.

UV-C tower for point-of-care decontamination of filtering facepiece respirators.

BACKGROUND: Filtering facepiece respirators (FFR) are critical for protecting essential personnel and limiting the spread of disease. Due to the current COVID-19 pandemic, FFR supplies are dwindling in many health systems, necessitating re-use of potentially contaminated FFR. Multiple decontamination solutions have been developed to meet this pressing need, including systems designed for bulk decontamination of FFR using vaporous hydrogen peroxide or ultraviolet-C (UV-C) radiation. However, the large scale on which these devices operate may not be logistically practical for small or rural health care settings or for ad hoc use at points-of-care. METHODS: Here, we present the Synchronous UV Decontamination System, a novel device for rapidly deployable, point-of-care decontamination using UV-C germicidal irradiation. We designed a compact, easy-to-use device capable of delivering over 2 J cm2 of UV-C radiation in one minute. RESULTS: We experimentally tested Synchronous UV Decontamination System' microbicidal capacity and found that it eliminates near all virus from the surface of tested FFRs, with less efficacy against pathogens embedded in the inner layers of the masks. CONCLUSIONS: This short decontamination time should enable care-providers to incorporate decontamination of FFR into a normal donning and doffing routine following patient encounters.