Dataset Information

Rapid antibiotic susceptibility testing based on bacterial motion patterns with long short-term memory neural networks.

ABSTRACT: Antibiotic resistance is an increasing public health threat. To combat it, a fast method to determine the antibiotic susceptibility of infecting pathogens is required. Here we present an optical imaging-based method to track the motion of single bacterial cells and generate a model to classify active and inactive cells based on the motion patterns of the individual cells. The model includes an image-processing algorithm to segment individual bacterial cells and track the motion of the cells over time, and a deep learning algorithm (Long Short-Term Memory network) to learn and determine if a bacterial cell is active or inactive. By applying the model to human urine specimens spiked with an Escherichia coli lab strain, we show that the method can accurately perform antibiotic susceptibility testing as fast as 30 minutes for five commonly used antibiotics.

SUBMITTER: Iriya R

PROVIDER: S-EPMC7241544 | biostudies-literature | 2020 May

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Rapid antibiotic susceptibility testing based on bacterial motion patterns with long short-term memory neural networks.

Iriya Rafael R Jing Wenwen W Syal Karan K Mo Manni M Chen Chao C Yu Hui H Haydel Shelley E SE Wang Shaopeng S Tao Nongjian N

IEEE sensors journal 20200117 9

Antibiotic resistance is an increasing public health threat. To combat it, a fast method to determine the antibiotic susceptibility of infecting pathogens is required. Here we present an optical imaging-based method to track the motion of single bacterial cells and generate a model to classify active and inactive cells based on the motion patterns of the individual cells. The model includes an image-processing algorithm to segment individual bacterial cells and track the motion of the cells over ...[more]

PMID: 32440258

Similar Datasets

Project description:Flow cytometry holds promise to accelerate antibiotic susceptibility determinations; however, without robust multidimensional statistical analysis, general discrimination criteria have remained elusive. In this study, a new statistical method, probability binning signature quadratic form (PB-sQF), was developed and applied to analyze flow cytometric data of bacterial responses to antibiotic exposure. Both sensitive lab strains (Escherichia coli and Pseudomonas aeruginosa) and a multidrug resistant, clinically isolated strain (E. coli) were incubated with the bacteria-targeted dye, maltohexaose-conjugated IR786, and each of many bactericidal or bacteriostatic antibiotics to identify changes induced around corresponding minimum inhibition concentrations (MIC). The antibiotic-induced damages were monitored by flow cytometry after 1-h incubation through forward scatter, side scatter, and fluorescence channels. The 3-dimensional differences between the flow cytometric data of the no-antibiotic treated bacteria and the antibiotic-treated bacteria were characterized by PB-sQF into a 1-dimensional linear distance. A 99% confidence level was established by statistical bootstrapping for each antibiotic-bacteria pair. For the susceptible E. coli strain, statistically significant increments from this 99% confidence level were observed from 1/16x MIC to 1x MIC for all the antibiotics. The same increments were recorded for P. aeruginosa, which has been reported to cause difficulty in flow-based viability tests. For the multidrug resistant E. coli, significant distances from control samples were observed only when an effective antibiotic treatment was utilized. Our results suggest that a rapid and robust antimicrobial susceptibility test (AST) can be constructed by statistically characterizing the differences between sample and control flow cytometric populations, even in a label-free scheme with scattered light alone. These distances vs paired controls coupled with rigorous statistical confidence limits offer a new path toward investigating initial biological responses, screening for drugs, and shortening time to result in antimicrobial sensitivity testing.

Dataset Information

Rapid antibiotic susceptibility testing based on bacterial motion patterns with long short-term memory neural networks.

Publications

Rapid antibiotic susceptibility testing based on bacterial motion patterns with long short-term memory neural networks.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets