README File for Title: "Epifluorescence-based three-dimensional traction force microscopy" Authors: Lauren Hazlett(1,5,*), Alexander K. Landauer(2,5,*), Mohak Patel(2), Hadley Witt(3,4), Jin Yang(5), Jonathan S. Reichner(4), and Christian Franck(5,+) Affiliations: 1. Center for Biomedical Engineering, Brown University 2. School of Engineering, Brown University 3. Pathobiology Graduate Program, Brown University 4. Department of Surgery, Rhode Island Hospital 5. Mechanical Engineering, University of Wisconsin - Madison * Contributed equally + Corresponding Abstract: We introduce a novel method to compute three-dimensional (3D) displacements and both in-plane and out-of-plane tractions on nominally planar transparent materials using standard epifluorescence microscopy. Despite the importance of out-of-plane components to fully understand cell behavior, epifluorescence images are generally not used for 3D traction force microscopy (TFM) experiments due to limitations in spatial resolution and measuring out-of-plane motion. To extend an epifluorescence-based technique to 3D, we employ a topology-based single particle tracking algorithm to reconstruct high spatial-frequency 3D motion fields from densely seeded single-particle layer images. Using an open-source finite element (FE) based solver, we then compute the 3D full-field stress and strain and surface traction fields. We demonstrate this technique by measuring tractions generated by both single human neutrophils and multicellular monolayers of Madin-Darby canine kidney (MDCK) cells, highlighting its acuity in reconstructing both individual and collective cellular tractions. In summary, this represents a new, easily accessible method for calculating fully three-dimensional displacement and 3D surface tractions at high spatial frequency from epifluorescence images. We released and support the complete technique as a free and open-source code package. Funding: NIH/BMBI R01-AI116629-01 and direct support for H.W. via NIH/NIDCR F31DE02874 Note: Please see the associated publication for methods and imaging parameters. The entire code package can be found on the Franck Lab Github at: https://github.com/FranckLab. File Descriptions: Folder name: example_data Size: 247 MB Description: Contains synthetic practice data case. Contains 2 multipoints ('XYpoint_name_1' and 'XYpoint_name_2') and 3 time points ('000', '001', and '002') and an associated black-and-white binarized "cell" outline (exampleBWfile.mat) of synthetically generated data in the correct folder structure as a practice and example data set. Folder name: rigid_disp_data and rigid_disp_data_deconv Size: 12.0 GB and 18.2 GB Description: Contains all data required to replicate findings pertaining to Figure 2, separated into two folders due to large file size. Multipoint names pertain to the example rigid displacement cases (i.e., 'zerodisp' is the zero rigid displacement case, 'xdisp' is rigid motion prescribed in x-direction, 'ydisp' in the y-direction, and 'zdisp' in the z-direction). For the zero rigid displacement case, there are two timepoints: timepoint '_000' and '_001' are both experimental images captured at the same xyz location. In the case of 'xdisp', 'ydisp', and 'zdisp' multipoints, the timepoint data relates to amount of applied displacement as follows: '_000' is the reference image '_001' is 0 um displacement '_002' is 1 um '_003' is 2 um '_004' is 3 um '_005' is 5 um '_006' is 10 um '_007' is 15 um Note: for the figures, all data was run in cumulative mode. Additionally, there is a line in the code intended to adjust for rigid drift during cell experiments (a common occurrence due to drift in the system over time), which needs to be commented out to accurately measure the intentionally-applied rigid displacements here: to reproduce our results, comment out Line 85 in 'funRunTPT.m' if running in incremental mode, or Line 100 if running cumulative mode. Folder name: synthetic_validation_data Size: 1.89 GB Description: Contains data and post-processing scripts required to replicate findings pertaining to Figure 3. Data is synthetically generated (see the paper for a description), with multipoint names 'force_scaling', and 'width_scaling'. Also contained are the output files from Matlab for FEniCS, contained in the 'matlab_results' subfolder, the analytical applied fields are contained in the subfolder 'ground_truth', and all additional scripts are contained in 'analysis_scripts'. Note: these files were run in cumulative mode for the figures presented in the publication. Folder name: cell_data Size: 3.53 GB Description: Contains all data required to replicate findings pertaining to Figure 4, plus one extra timepoint for each multipoint (i.e., MDCK and neutrophil) to help users familiarize themselves with cell displacement and traction plots. In total, there are two multipoints 'MDCK' and 'neutrophil', with 3 timepoints each ('_000', '_001', and '_002'), although these timepoints cannot be run simultaneously through the code due to the differences in imaging parameters between the data sets (these are not statistically comparable image sets). All of the outputs from running these files through the code are included, with outputs for and from FEniCS placed in a separate subfolder 'FEniCSdata'. Note: these files were run in cumulative mode for the figures presented in the publication.