When extracting experimental data from a SOFT formatted family file using the Import Samples wizard in the Subio Platform, information is automatically populated into the Sample Information. At this time, the Sample_characteristics_ch1 field contains all experimental parameters concatenated using the delimiter “///”. However, when importing a SOFT formatted family file via the Look Up tool in Edit Parameters or Edit Sample Information, only the last parameter is imported from Sample_characteristics_ch1, discarding all other values.
To work around this issue, I'll introduce a method to easily extract experimental parameters by having ChatGPT generate Python code and executing it in JupyterLab. While outputting a table of experimental parameters is convenient for importing into Edit Parameters, importing such data into Sample Information can cause columns to explode exponentially, potentially making it unmanageable. Therefore, we will also introduce a method for importing into Sample Information and a method for utilizing the imported information in Edit Parameters as experimental parameters.
Getting Experiment Parameters form The Characteristics Field of GEO Samples
In the GEO's sample records, experiment parameters are often stored in the Characteristics field, which allows authors to store pairs of titles and values to describe their studies. But it might be a bit hard to collect them from all the samples if you don't have coding skills. This movie illustrates how to use ChatGPT and Jupyter to perform the task and import the results into the Subio Platform for dataset analysis.