Skip to Main content Skip to Navigation

Automation of AFM measurements for biological applications

Sergio Proa Coronado 1
1 LAAS-ELIA - Équipe Ingénierie pour les sciences du vivant
LAAS - Laboratoire d'analyse et d'architecture des systèmes
Abstract : In recent years mechanical phenotype of cells (such as adhesion, elasticity, stiffness), also known as mechanical properties, has been proved to be a valid identifier to distinguish healthy cells from diseased cells. Researchers around the world have related the mechanical phenotype to cancer, cardiovascular, and blood-related diseases, among others. Atomic force microscopy (AFM) is the most used technique to measure mechanical properties, and it makes it possible to obtain the properties at the nanoscale. However, AFM manipulation requires high technical skills, and initially, it was not designed for biological samples. However, the capability to analyze samples in air or liquid in recent years makes it more appealing for the living area. However, using the current AFMs, it is not straightfoward to analyze a high number of cells or cell population and then it is difficult to obtain statistical results. This doctoral thesis aims at solving the problem of low throughput, an automated methodology is proposed to do it. This methodology is based on the combination of two techniques, cell arrays, and AFM automation. The mechanical measurements are done automatically by executing a developed Jython script, and the cells are immobilized in known positions proposing here a number a number of conducted measurements compared with was found in the literature. Firstly, Immobilization is done for the microbes in microfabricated PDMS stamps and the mammalian cells in commercial cell arrays (BIOSOFT and CYTOO). The immobilization is done, in the case of the microbes, using convective/capillary assembly technique reaching ~85 % filling rate. And for the mammalian cells, the technique used attached the cell to a surface previously functionalized. Next, the AFM was modified to perform the measurements automatically. The automation was made by developing a Jython written script and executed directly in a commercial BioAFM (JPK Germany). The script is versatile, and it has been adapted, or it can be adapted to several sample configurations. The script performs a small number of indentations (9 or 16) on the sample, acquiring force curves from different regions of the cells and at the same time, reducing the time spent on each cell. The results demonstrated that increasing the number of cells impacts the number of measurements done to the cells, and it is still possible to obtain results comparable to the results reported in the literature. For the first time, the stiffness analysis on ~900 yeast cells (C. albicans) is reported, and it is evident the presence of two subpopulations. We compared native yeast cells with caspofungin treated yeast cells. The results were obtained in 4 h with 9 indentations per cell. For the HeLa cells, the comparison was made between native HeLa cells and fixed HeLa cells; and ~80 cells were analyzed in 30 minutes. In both cases (mammalian and yeast cells), a shift between the native and treated cells was observed, this shift agrees with the literature and proves that is possible to reduce the number of indentations done to the cells if the number of analyzed cells is high. Thanks to the massive amount of data collected, it was possible to use Machine learning, and the preliminary results show that it is possible to distinguish between native and treated cells. The differentiation was made entering the descriptors: stiffness, adhesion, and work of adhesion to the machine learning algorithm. This work contributes to achieving a statistical significance, which is one of the main drawbacks of AFM mechanical analysis and can be considered as one of the first steps to have a diagnostic tool from the atomic force microscope.
Complete list of metadata
Contributor : Abes Star :  Contact
Submitted on : Tuesday, November 24, 2020 - 4:56:10 PM
Last modification on : Thursday, June 10, 2021 - 3:03:56 AM


Version validated by the jury (STAR)


  • HAL Id : tel-02943902, version 2


Sergio Proa Coronado. Automation of AFM measurements for biological applications. Biological Physics []. Université Paul Sabatier - Toulouse III; Instituto politécnico nacional (México), 2020. English. ⟨NNT : 2020TOU30062⟩. ⟨tel-02943902v2⟩



Record views


Files downloads