Phathom is a Python package for analyzing terabyte-scale biological images. It relies on distributed computing to scale up image processing tasks that are commonly encountered when using Selective Plane Illumination Microscopy (SPIM). These image processing task include:

• Non-rigid registration for multi-round imaging
• Individual nuclei segmentation
• Cell phenotype classification
• Brain atlas alignment

Clone this repository and run python setup.py install. Documentation is hosted on readthedocs.

The purpose of Phathom is to improve the scalability of computational pipelines used in biological image processing. Popular open-source tools such as Scikit-image cannot directly process images that are larger than memory. To address this, Phathom uses persistent Zarr arrays to store and access large images on-disk or from a shared file system. For processing, Phathom uses the SCOOP library to coordinate a pool of workers, either locally or in a distributed manner. Phathom supports HPC environments using SLURM.

Phathom is released as an open-source package on October, 2021, and maintained by members of the Kwanghun Chung Lab at MIT.

• Original Author: Justin Swaney
• Primary Contact: Lee Kamensky ([email protected])