Machine learning method for extracting elastic modulus of cells.

The Hertz contact mechanics model is commonly used to extract the elastic modulus of the cell, but the basic assumptions of the model are often not met in cell indentation experiments, which can lead to errors in the obtained elastic modulus of cell. The establishment of theoretical formulas or modification of the Hertz formulas has been proposed to reduce the errors introduced by indentation depth and cell thickness, but errors from cell radius and probe radius are largely neglected. Herein, we build a neural network model in machine learning to extract the elastic modulus of cell, which takes into account of four variables: indentation depth, cell thickness, cell radius, and probe radius. The validity of the model is demonstrated by the indentation experiment. The introduction of machine learning methods provides an alternative solution for extracting the elastic modulus of the cell and has potential for application.