Identifying Virulence Determinants In Pathogenic Mycobacteria Via Changes In Host Cell Mitochondrial Morphology

TL;DR

This study develops a computational model to analyze mitochondrial morphological changes caused by pathogenic mycobacteria, aiming to identify bacterial virulence factors through image processing and machine learning.

Contribution

It introduces a novel graph-based statistical approach to recognize mitochondrial phenotypes linked to mycobacterial infection, enabling large-scale virulence screening.

Findings

Mitochondrial morphology changes can be quantitatively characterized.

A graph-based statistical method effectively identifies infection-induced phenotypes.

Potential for high-throughput virulence gene screening in mycobacteria.

Abstract

The goal of this study is to develop a computational model of the progression of changes in mitochondrial phenotype resulting from infection with pathogenic mycobacteria. This ultimately will enable a large-scale virulence screen of mutant bacterial libraries. Mycobacterium tuberculosis (Mtb) is an intracellular pathogen, but only a small number of its genes have been studied for roles in intracellular host cell survival and replication. Mitochondria are the powerhouse of the host cell and play critical roles in cell survival when attacked by certain pathogens. When Mtb bacteria invade host cells, they induce changes in mitochondrial morphology, making mitochondria a novel target for image processing and machine learning to determine virulence associations of genes in Mtb and potentially other related intracellular pathogens. By hypothesizing mitochondria as an instance of a dynamic and interconnected graph, we demonstrate a statistical approach for quantitatively recognizing novel mitochondrial phenotypes induced by invading pathogens.