An automated and reproducible workflow for running and analyzing neural simulations using Lancet and IPython Notebook

Area of Science:

• Computational Science
• Scientific Computing
• Reproducible Research

Background:

• Managing complex computational tasks and ensuring reproducibility in scientific research is challenging.
• Existing workflows may lack flexibility or constrain scientific exploration.

Purpose of the Study:

• To introduce Lancet, a Python utility for managing large batches of computational runs.
• To demonstrate a workflow combining Lancet with IPython Notebook for reproducible scientific research.

Main Methods:

• Lancet provides a simulator-independent way to specify, launch, and collate results from demanding program runs.
• IPython Notebook captures computational steps, automating and documenting the research process.
• The approach supports any program launched as a process with file output.

Main Results:

• Lancet facilitates fully reproducible scientific research through automated documentation of computational processes.
• The combination of Lancet and IPython Notebook offers a flexible and agile workflow.
• The tool has been successfully applied in computational neuroscience and microprocessor simulations.

Conclusions:

• Lancet offers a general and adaptable solution for managing complex computational research.
• This approach enhances the reproducibility and transparency of scientific findings.
• The utility supports evolving research needs and tools over time.