Collaborative Research: ABI Innovation:Large-scale Analysis of the Evolution of Organellar Networks

University Of Pittsburgh, Pittsburgh PA

Investigators

Abstract

This project will design and implement a high-throughput, large-scale computational pipeline to identify and track changes in organellar shape, quantity, and spatial distribution over large sequences of Z-stack microscope images and digital videos, and provide a causal and/or non-causal interaction map of how the subcellular structures evolve individually and as a group. Two well-studied intracellular bacterial pathogens, Mycobacterium tuberculosis (Mtb) and Legionella pneumophila (Lp), known to induce dramatic organellar changes in host cells they infect, will be used to test the framework. Accurately characterizing the spatio-temporal evolution of the organellar changes in a large collection of time-lapse videos from different genetic mutants and controls will provide significant insight into the host response as a function of the specific stimulus. This can have a positive impact on large-scale genotype-phenotype association studies and the efficient screening of large genomic libraries. The analytical framework will be implemented using open source tools and integrated with standard tools available to the imaging community, thus increasing the accessibility of these tools and the transparency of the analysis. Open source development will facilitate enrichment of undergraduate education, through synergies with an NSF and Department of Defense funded REU program, TECBio. The problem of isolating and tracking spatiotemporal changes in individual organelles is extremely challenging, subject to significant sources of noise and variability under many possible conditions. This project relies on a novel approach of considering subcellular components as vertices in a social network. Ensembles of cellular machinery will be characterized as networks to study organellar response to stimuli within the framework of interconnected elements whose collective behavior influence local and global social network topologies. The team will develop spectral graph analytics to characterize global network evolution and identify local dynamics, thus providing powerful algorithmic tools for isolating cohesive regions of subcellular evolution. This framework will be designed from the ground up to be highly scalable, capable of processing arbitrary amounts of high-resolution 4D confocal microscopy image data. Open-source, distributed software for "big data" analysis, such as Apache Spark and Hadoop, will be used to build the framework. The framework will be developed and released under an open source license, providing biology researchers with a powerful tool for performing large-scale spatiotemporal condition-response studies. The results of this study will be disseminated from: http://www.csb.pitt.edu/faculty/chakra/.