Colorado State University
Collaborators: Jiangyu Sun, Colorado State University; Lubna Tahtamouni, Colorado State University and The Hashemite University; Ashok Prasad, Colorado State University
Title: Linking Cell Morphology to Cell Identity Using Machine Learning
Short Abstract: Cell morphology and cytoskeletal organization can provide great insight into cell health and disease states, as the cytoskeleton plays essential roles in countless cell processes, from cell division to migration to signaling. In cancer cells, cytoskeletal dynamics, cytoskeletal filament organization, and overall cell morphology are known to be altered substantially. This project used a small fluorescence microscopy image dataset of retinal pigment epithelial (RPE) cells to investigate the effectiveness of convolutional neural networks (CNNs) to distinguish between normal and oncogenically transformed cells and between different subtypes of transformed cells of the same cell line. We found that cell morphology was a sensitive signature of cell identity, and thus could be a very useful method for assaying cell phenotype.
Link to Full Abstract: Ashok Prasad