PyCurv

PyCurv estimates membrane curvature from volumetric and surface segmentations to quantify local membrane shape and morphology for structural and cellular biology analyses.

Key Features:

• Membrane curvature estimation: Estimates local membrane curvature from volumetric and surface segmentations, enabling analysis at molecular resolution.
• Handling artifacts: Manages quantization noise and open borders common in cryo-electron tomography (cryo-ET) image acquisition and membrane segmentation.
• Data compatibility: Supports volumetric formats MRC, EM, VTI, NII and surface formats VTP, VTK, STL, PLY.
• Signed surface extraction: Extracts a signed surface as a triangle mesh from membrane segmentations.
• Graph-based representation: Converts the triangle mesh into a graph to identify neighboring triangles and compute geodesic distances for local analyses.
• Tensor voting algorithm: Applies tensor voting to estimate curvature while providing robust estimations of surface normals and principal directions.
• Distance and thickness measurements: Computes distances between adjacent membranes and organelle thicknesses from segmentation-derived surfaces.
• Versatility across imaging modalities: Operates on cryo-electron tomography (cryo-ET), confocal light microscopy, and MRI volumetric data.

Scientific Applications:

• Membrane morphology analysis: Quantifies local membrane shape to study cellular ultrastructure and membrane remodeling at near-native states and molecular resolution.
• Organelle structural measurements: Measures organelle thicknesses and inter-membrane distances for morphological and comparative studies.
• Structural biology support: Provides curvature and directional information useful for interpreting cryo-ET and light microscopy data in structural biology investigations.

Methodology:

From membrane segmentation a signed triangle mesh is extracted, the mesh is converted to a graph to identify neighboring triangles and compute geodesic distances, and tensor voting is applied to estimate curvature, surface normals, and principal directions.