2_3D_cell_nuclear_shape_modeling_morphometry

SOCR 3D Cell Morphometry Project

2. 3D shape modeling and morphometry

Documentation for 3D surface modeling and morphometric features extraction

Inputs: individual c0 and c2 binary masks in TIFF format (available for downloading on project web-page).

Nuclei and nucleoli 3D shape modeling and morphometry (both DAPI, c0, and EtBr, c2):

This protocol uses parts of LONI Global Shape Analysis (GSA) Workflow:

1. Convert binary masks into NIfTY format
2. Model surface of each binary mask as a 2-manifold embedded in 3D space and represented as triangulated mesh using Mask2Mesh [1], followed by mesh translation, simplification, and subdivision using ShapeTools [2], and extraction of 6 morphometric measures using ShapeReporter [2]
3. Build nucleus feature vector by concatenating nuclear morphometric measures with nucleolar measure sample statistics grouped by nucleus (e.g., average, minimum, maximum, etc.) and perform model training and prediction using scikit-learn package [3]

Outputs: classification performance metrics (accuracy, precision, recall, AUC)

Required 3rd-party software:

[1] MOCA

[2] LONI ShapeTools

[3] scikit-learn