A Deep Learning Pipeline for Segmentation of Proteus Mirabilis Colony Patterns
Video
Team Information
Team Members
Anjali Doshi, PhD Candidate, Department of Biomedical Engineering, Columbia Engineering
Marian Shaw, PhD Student, Department of Biomedical Engineering, Columbia Engineering
Ruxandra Tonea, Undergraduate Student, Department of Biomedical Engineering, Columbia Engineering
Rosalía Minyety, Undergraduate Student, Department of Biomedical Engineering, Columbia Engineering
Soonhee Moon, Research Associate/Visual Researcher, School of Engineering and Applied Science, Columbia University, Department of Biomedical Engineering
Faculty Advisors:
Andrew Laine, Professor, School of Engineering and Applied Science, Columbia University, Department of Biomedical Engineering
Jia Guo, Assistant Professor, Mortimer B. Zuckerman Mind Brain Behavior Institute, Columbia University Medical Center, Department of Psychiatry (CUMC)
Tal Danino, Associate Professor of Biomedical Engineering, Columbia Engineering
Abstract
The motility mechanisms of microorganisms are critical virulence factors, enabling their spread and survival during infection. Motility is frequently characterized by qualitative analysis of macroscopic colonies, yet the standard quantification method has mainly been limited to manual measurement. Recent studies have applied deep learning for classification and segmentation of specific microbial species in microscopic images, but less work has focused on macroscopic colony analysis. Here, we advance computational tools for analyzing colonies of Proteus mirabilis, a bacterium that produces a macroscopic bullseye-like pattern via periodic swarming, a process implicated in its virulence. We present a dual-task pipeline for segmenting (1) the macroscopic colony including faint outer swarm rings, and (2) internal ring boundaries, unique features of oscillatory swarming. Our convolutional neural network for patch-based colony segmentation and U-Net with a VGG-11 encoder for ring boundary segmentation achieved test Dice scores of 93.28% and 83.24%, respectively. The predicted masks at times improved on the ground truths from our automated annotation algorithms. We demonstrate how application of our pipeline to a typical swarming assay enables ease of colony analysis and precise measurements of more complex pattern features than those which have been historically quantified. An implementation of our work can be found on https://github.com/daninolab/proteus-mirabilis.
Team Lead Contact
Anjali Doshi: apd2136@columbia.edu
