Integration of SPIDER workflows into Scipion for user-friendly analysis of electron microscopy data

T. R. Shaikh1, J. M. de la Rosa-Trevín2, R. Marabini3, J. M. Carazo2

1Central European Institute of Technology, Masaryk University, 625 00 Brno, Czech Republic

2Centro Nacional de Biotecnologia–CSIC, Campus Canto Blanco, 28049 Madrid, Spain

3Escuela Politécnica Superior, Universidad Autónoma de Madrid, Campus Universidad Autónoma, 28049 Cantoblanco, Madrid, Spain

 

In three-dimensional electron microscopy, there are 7 to 10 general-purpose image-processing software packages, and a few dozen other packages which perform specific functions or upon specific types of samples.  It can be useful to mix and match the best features of the different software packages, but to do so, data historically often had to be converted from one format to another, which presented a barrier with regard to interoperability.

 

Scipion is a workflow integration system, among the functions of which is to manage image-processing projects, consolidate access to different software suites, trace the inputs and outputs of each step, monitor processes, etc.  SPIDER is one of the oldest image-processing suites currently in use, available since 1978, with over 600 functions and its own internal scripting language.  The goal of the project we are presenting is to provide a framework to integrate the functionalities of SPIDER, at the procedural level, into the Scipion environment.  Some advantages of Scipion to manage SPIDER workflows are as follows.  First, display of outputs can be customized according to the data type, e.g., an image, a set of images, a 2D plot, a volume, etc.  Second, inputs and outputs are classified into data types, and subsequent procedures will list, as possible options for an input, existing files of the appropriate data type.  Third, some parameters can be tested interactively before execution by using “wizards” in real time, rather than by trial & error. 

 

As a case study, the SPIDER workflow for 2D multivariate data analysis was integrated into Scipion.  Of particular importance is the method to parse SPIDER procedures from Scipion to retrieve  the necessary information about the script parameters.  The next tasks will be integration of SPIDER's random conical tilt and projection-matching workflows, two of the workhorses of SPIDER usage.