(1254-C) Deciphering Cellular Mechanisms through Phenotype: CRISPR-Induced Perturbations Analyzed through Cell Painting and Machine Learning
Tuesday, February 6, 2024
12:00 PM – 1:00 PM EST
Location: Exhibit Halls AB
Abstract: This study explores complex biological systems by employing CRISPR-induced perturbations, analyzed through a phenotypic screening experiment, to decode biological patterns via quantitative morphology analysis. At its core is the Cell Painting microscopy assay, which captures detailed behaviors of various organelles to map gene function with respective morphology. We used the open-access JUMP-Cell Painting CRISPR dataset, featuring 7975 unique gene knock-outs; totaling to over 40k+ replicates, providing readouts from 6 dyes across 5 imaging channels and 8 cellular components in the U2OS cell line. Customized machine learning and bioinformatics methods interpreted the vast dataset, revealing alignments between genetic perturbations and cellular mechanisms. Through similarity and anti-similarity analysis, we identified morphological neighbors, suggesting potential interactors for proteins of interest. This aspect offers a new perspective in the understanding of protein interactions, highlighting the predictive power of our approach - aligned with cell-level effect rather than gene-level quantitative balance as in omics approaches. Our findings not only replicate insights from the literature but also unveil novel interdependencies within cellular machinery. This underscores the phenotypic assay's advantage in capturing subtle yet crucial interactions. Integrating CRISPR technology with advanced imaging and analysis, our research offers a nuanced view of genetic alterations' impact on cellular morphology to understand cellular mechanisms and guide therapeutic strategies, demonstrating the synergy between various modalities in revealing the complexities of biological systems.