This repository contains tables and jupyter notebooks needed to reproduce analysis and generate all plots from main and supplementary figures from the article "Single-cell bacterial susceptibility to bacteriostatic and bactericidal antibiotics and its relation to small and large population statistics" by Maikranz et al. Of particular interest might be "InferenceValidation.ipynb" notebook, which allows to simulate the performance of the inference techniques for various parameters, which includes Poisson and negative binomial distributon of initial cell number.
Understanding and countering the global rise of antibiotic resistance requires insights into the heterogeneity of antibiotic responses within a single bacterial population through single-cell measurements. While such measurements are enabled by microfluidic droplets, the interpretation of the droplet-level counting in terms of single-cell susceptibility and its translation on the population-scale are lacking a theoretical foundation. Here we develop a theoretical framework to analyze droplet measurements based on paired observations of the initial cell count and final outcome within each droplet, while accounting for experimental uncertainties. The model is applied to obtain single-cell susceptibility measurements on hundreds of experiments with four different antibiotics. The cells respond to bactericidal antibiotics in a reproducible and sharp manner, while the response to bacteriostatic molecules is more variable and takes place over a wider range of concentrations. These results provide a microscopic interpretation of population-scale measurements.