When importing experimental data from a SOFT formatted family file in Subio Platform, sample metadata is automatically populated into Sample Information.
In GEO datasets, experimental parameters are often stored in the Sample_characteristics_ch1 field, where multiple parameters are concatenated using the delimiter “///”.
When using the Import Samples wizard, all parameters in this field are correctly imported.
However, in the current version of Subio Platform, when using the Look Up tool in Edit Parameters or Edit Sample Information, only the last parameter is imported from the Sample_characteristics_ch1 field, while the others are discarded.
This is a known issue in the current version.
In this article, we introduce a practical workaround.
Extracting Parameters Using ChatGPT and Python
To address this issue, we use Python code generated by ChatGPT and execute it in JupyterLab to extract all experimental parameters.
This approach allows you to:
- Extract all parameters from the Characteristics field
- Convert them into a structured table
- Import them into Subio Platform for analysis
Important Considerations
Due to this limitation, the original experimental parameters stored in the Sample_characteristics_ch1 field are not fully preserved when using the Look Up tool.
Working with GEO Sample Characteristics
In GEO sample records, experimental parameters are typically stored in the Characteristics field as key-value pairs.
Extracting these parameters across all samples can be challenging without coding.
This video demonstrates how to:
- Use ChatGPT to generate Python code
- Execute the code in JupyterLab
- Extract parameters from GEO samples
- Import the results into Subio Platform
Next Step
For a complete guide to RNA-Seq data analysis, please see the full tutorial:
→ RNA-Seq Data Analysis Tutorial
After preparing your dataset and parameters, you can follow the tutorial to proceed with downstream analysis and interpretation.