Active learning model for extracting elastic modulus of cell on substrate

Abstract

The cell elastic modulus (Ec) is widely used as the mechanics-based marker to analyze the biological effects of substrates on cells. However, the employment of the Hertz model to extract the apparent Ec can cause errors due to the disobedience of the small deformation assumption and the infinite half-space assumption, as well as an inability to deduct the deformation of the substrate. So far, no model can effectively solve the errors caused by the above-mentioned aspects simultaneously. In response to this, herein, we propose an active learning model to extract Ec. The numerical calculation with finite element suggests the good prediction accuracy of the model. The indentation experiments on both hydrogel and cell indicate that the established model can efficiently reduce the error caused by the method of extracting Ec. The application of this model may facilitate our understanding about the role of Ec in correlating the stiffness of substrate and the biological behavior of cell.