Computational methods for analyzing cell shape heterogeneity with application to Atomic Force Microscopy.

Cells cultured on planar surfaces adopt a variety of morphological shapes. Despite their seeming randomness, such morphology is determined by a biophysical process that represents the sum of multiple cellular processes. Cell morphology is dynamic, and cells change their shape with time and across different time scales, e.g. with local modification of the environment, or from differentiation/cancer. Morphological trajectories are likely to be even more informative than static images, but we lack basic tools to understand these trajectories. In this context, Atomic Force Microscopy is an advanced technique for investigating morphological properties of cell, by scanning the surface of a biological sample. Our proposed project is to develop mathematical methods to analyze morphological shape heterogeneity in the context of biological image data, and develop more specific methods to automate the analysis of images from Atomic Force Microscopy. These methods will lead to go beyond traditional approaches for clustering and features detection of cell images, to provide a more dynamic picture of the cell shape variability.

Faculty Supervisor:

Khanh Dao Duc

Student:

Partner:

CentraleSupélec

Discipline:

Computer science

Sector:

Education

University:

The University of British Columbia

Program: