EMAN2 and EMEN2: Flexible Python-based platforms for electron microscopy

Authors: Rees, Ian, Baylor College of Medicine; Ludtke, Steven, Baylor College of Medicine

Track: Bioinformatics

Three-dimensional electron microscopy has developed into a powerful technique for obtaining structural information that often cannot be obtained with other methods. Progress in the field continually produces new experimental techniques for sample preparation, imaging, image processing, reconstruction algorithms, and structural analysis tools, resulting in rich data sets with substantial amounts of experimental variation. This constantly evolving landscape requires flexible software platforms for developing new algorithms and sharing data with collaborators and the public. At the National Center for Macromolecular Imaging, we have developed two open-source, Python-based tools: EMAN2 and EMEN2.

EMAN2 is a modular, extensible scientific image processing suite implemented in Python and C++. This mixed approach allows speed-critical image processing algorithms to be written in C++, as high resolution 3D reconstructions can require tens of thousands of images and hundreds of thousands of CPU hours. The core distribution includes over 500 algorithms, and new algorithms can easily be added and called from C++ or Python. Conversely, application-level code and UIs are written in Python, which enhances developer productivity and lowers the level of difficulty for scientists to write new programs.

EMEN2 is an object-oriented scientific database and electron lab notebook, with a flexible schema based on user-created descriptions of experimental protocols. These descriptions allow investigators to quickly add new types of data to the system without requiring a database administrator. EMEN2 has a flexible security model for sharing data with collaborators and public. EMEN2 uses Berkeley DB for the storage backend, provides a web interface built using Twisted and MakoTemplates, and has full API access via JSON-RPC and XML-RPC. A desktop client, EMDash, for uploading raw data from instruments in real time.

While both EMAN2 and EMEN2 have their roots in the electron microscopy community, we hope they may be of interest to users in other domains.