Introduction

AIgarMIC is a python package designed to assist bacteriologists in automating the analysis of agar dilution minimum inhibitory concentration testing. At a high level, AIgarMIC is designed to take a set of images of agar plates, each containing multiple strains of bacteria exposed to a different antimicrobial concentration per plate, and return a minimum inhibitory concentration for each strain.

Functionality

AIgarMIC offers the following core functionality that allows users to convert a set of agar plate images to a list of minimum inhibitory concentrations:

  1. Sectioning of images into smaller images, each containing a bacterial colony (or lack thereof),

  2. Classification of each colony image into a growth quality category (e.g., no growth, inhibited growth, or good growth), using a pre-trained convolutional neural network classifier,

  3. Conversion of each colony image into a three dimensional matrix, where the first two dimensions represent the row and column positions of each colony inoculation, and the third dimension is the antimicrobial concentration,

  4. Algorithmic determination of the minimum inhibitory concentration for each strain, using the three-dimensional growth matrix,

  5. Determination of quality assurance metrics for each strain (e.g., presence of growth in the antimicrobial-negative control agar plate).

User interface

Users have two options for interacting with AIgarMIC:

  1. Use as a traditional python package, where users can import the package, use classes, and call functions directly, or

  2. Use as a command line interface application, without needing to write python code (see Command-line Interface).

In general, the first option is required if users want to customize AIgarMIC or use advanced configurations. The second option is recommended for users who want to quickly analyze a set of agar plate images using pre-trained models and default configurations.

Typical workflows

Use as an MIC calculator

AIgarMIC can be used as a simply agar dilution MIC calculator. This can be useful for users who do not want to set up imaging stations, and are happy to manually annotate growth on an agar plate, e.g., using a spreadsheet software. For more details and examples, see MIC calculation.

Use with pre-trained models to convert images to MICs

Most users will likely use AIgarMIC to convert images of agar plates to MICs using the pre-trained models described in the validation manuscript: http://dx.doi.org/10.1128/spectrum.04209-23

AIgarMIC provides a command-line interface to do this without needing to write python code. A description and example can be found in Quickstart guide.

Advanced usage

AIgarMIC allows extension and customisation. For an advanced example of this, see Advanced Usage.