This page contains a demo accompanying the paper:
3D Shape Modeling for Cell Nuclear Morphological Analysis and Classification
Alexandr A. Kalinin,
Ari Allyn-Feuer, Alex Ade,
Gordon-Victor Fon, Walter Meixner, David Dilworth, Syed S. Husain, Jeffrey R. de Wet, Gerald A. Higgins, Gen Zheng,
Amy Creekmore, John W. Wiley, James E. Verdone, Robert W. Veltri, Kenneth J. Pienta, Donald S. Coffey,
Brian D. Athey,
Ivo D. Dinov
Scientific Reports, 2018
This demo is prepared for classification of serum-starved Fibroblast cells ( SS, #160 ). This workflow take as an input original 16 1024x1024xZ 3D TIFF images (sub-volumes) in DAPI channel (c0) and metadata. It demonstrates nuclear binary mask preparation, 3D shape modeling, morphometric measure extraction, and classification running in distributed mode on a cluster using LONI Pipeline guest mode. It outputs .csv file with image-level output label, nucleus-level accuracy and average probability as well as labels and probabilities for individual nuclear masks that were segmented out of 3D input sub-volume, passed the curation, 3D shape modeling, feature extraction, and classification.
You can double-click on group in the workflow at any moment to see individual modules inside. You can disconnect while the workflow is running – under Connections you will be able to see your unique GUEST-ID that you can use to reconnect later and check workflow status (enter space for password). Having your GUEST-ID you should be able to use LONI Pipeline Web App to reconnect to the same sessions (web app is still in Beta and might not work as expected). Workflow protocol can be ran multiple times to validate reproducibility of the morphometry results. Pipeline documentation, including instructions module definition, modification, and execution, is available on the official website.