ClueGO: GO term duplication

[Laika]

Hi,

I am new to ClueGO and attempting to use it for downstream analysis for WGCNA. I import the list of genes from a certain module (genes are clustered into a module based on similar expression patterns across samples). I used all the GO databases and KEGG for tribolium castaneum, default setting (GO term level 3-8, also GO fusion, pathway with pV <0.5 etc.) but sometimes I got networks with duplicated GO terms (e.g. "Lysosome" and "lysosome"). I would really appreciate some tips on picking reasonable parameters for such analysis.

Many thanks,

Laika

[Alex Pico]

Hi. I would normally suggest reaching out to the ClueGO app authors. Their email contact info is on the app store page for ClueGO.  But, I'm also testing out a new Cytoscape Helper GPT, so I'll share their response as well.

Best,

 Alex

------

Testing a GPT trained on Cytoscape user documentation, tutorials and previously answered questions. Feel free to ignore. Your questions will continue to be answered by the Cytoscape team on this platform. But any feedback you have on the quality of these responses would be appreciated.

------

Hello Laika,

It's great to hear that you're using ClueGO for your WGCNA downstream analysis. ClueGO is a powerful tool for analyzing gene sets in the context of biological pathways and networks, but it does require careful parameter setting to get meaningful results. Here are some tips for setting parameters in ClueGO, especially when dealing with issues like duplicated GO terms:

1. Gene Ontology (GO) Term Levels: Adjusting the GO term levels (e.g., 3-8) can help in focusing your analysis on a certain level of specificity. Lower levels (closer to 1) are more general, while higher levels are more specific. If you're getting duplicated terms, try adjusting these levels to either focus on more general or more specific terms.

2. Evidence Code Filter: Sometimes, the inclusion of certain evidence codes can lead to redundancy or less reliable terms. Check the evidence codes you're including in your analysis. For instance, excluding IEA (Inferred from Electronic Annotation) might reduce redundancy and improve the reliability of your results.

3. GO Fusion and Grouping: ClueGO has a feature called GO Fusion which combines related terms. While this can be useful to reduce clutter, sometimes it might merge terms that you would prefer to analyze separately. Adjusting this setting can help in managing the specificity of the terms.

4. Minimum and Maximum GO Level: These settings control the depth of the GO terms included. Adjusting these might help in filtering out redundant or less relevant terms.

5. P-Value Threshold: You mentioned using a p-value of <0.5, which is quite high and can include many terms that might not be statistically significant. Lowering this threshold (e.g., <0.05 or <0.01) can help in focusing on more relevant terms.

6. Taxon-Specific Analysis: Since you're working with Tribolium castaneum, ensure that the GO and KEGG databases you're using are specific to this organism, as this can impact the relevance of the GO terms identified.

7. Cluster Merging and Kappa Score: Adjust the kappa score threshold to control the stringency of term connections and clustering. This can also help reduce redundancy.

8. Manual Review: Sometimes, manual intervention is necessary to remove redundant terms or to interpret overlapping categories.

9. Updates and Versions: Ensure you are using the latest version of ClueGO and the GO/KEGG databases, as updates can fix bugs and include more accurate data.

For more detailed information and guidance on using ClueGO, you can refer to the ClueGO manual and documentation available on the ClueGO page in the Cytoscape App Store: ClueGO on Cytoscape App Store.

Remember, the choice of parameters can greatly influence your results, so it might be worth running multiple analyses with different settings to compare outcomes. Also, consider reaching out to the ClueGO community or forums if you encounter specific challenges or need more tailored advice.

Best regards, Cytoscape Helper