Option I – Virtual assay

What if I have analyzed data in a way that is not perfectly reusable or reproducible – for instance when I use a proprietary software or a colleague helped with data analysis – but I would still want to store and annotate the data analysis in a way comprehensible to others?

A computational workflow is like a protocol

Just like laboratory workflows produce samples or data, computational workflows produce data. They represent the processing steps used in data analysis or data transformation.

Create a virtual assay

A computational workflow can simply be treated as a protocol. This can be a summary of the steps followed in any data analysis software. Or it can be a script. Hence the data analysis can simply be packaged as an assay, with the computational workflow stored in protocols and the resulting data stored in dataset

1. Add a new assay Visualization.

2. Import the python script heatmap.py into protocols.

3. Import the figure heatmap.svg resulting from the script into dataset.

   • Directoryassays

     • DirectoryVisualization

       • Directorydataset

         • heatmap.svg
       • Directoryprotocols

         • heatmap.py
       • isa.assay.xlsx
       • README.md
   • …

4. Add a new annotation table to Visualization and click Create basic table!.

5. This will create the following building blocks:

   • Input [Sample Name]: Use File Picker to reference the proteomics output file.
   • Protocol Uri: Use File Picker to reference python script.
   • Output [Sample Name]: Use File Picker to reference to heatmap result file.

6. Rightclick then Edit to update the input and output columns to be Input [Data] and Output [Data] respectively.

   Input [Data]	Protocol Uri	Output [Data]
   sugar_result.csv	heatmap.py	heatmap.svg