4
$\begingroup$

I have recently started working on a substance's effect on a cell line in different dosages. for this, there is a tool called bmdexpress2 that I am using. Its input is the normalized counts from RNASeq for each dosage as a big matrix. When it comes to the pathway analysis step, unlike hallmarks of GSEA, this tool uses some databases for defining pathways which involve pathways of even 2-3 genes.

So what I want to discuss here is the strength of small pathways and filtering thresholds. How can we bioinformatically decide on the importance of a pathway made of 3 genes? Should we just filter them out as there are much more comprehensive pathways? Or do their GO Levels matter? Also there are many cases that out of those 3 genes, 1 of them is differentially expressed; is having 1 gene diff. expressed out of X genes in a pathway enough to say that the pathway is enriched?

Here is an example GO term for a small pathway.

And here is an example of the output I get from the tool. all genes platform column is the number of genes involved in that pathway, input genes column is how many of the genes in that pathway are significantly differentiated. I believe the rest is self explanatory.

enter image description here

Looking at the table, can we really say, ESCRT III Complex, for example is a good pathway to focus while having 1 gene out of 13 in a pathway but that change is significant?

Thanks

$\endgroup$
  • $\begingroup$ Why is the GO term relevant? Is that one of the terms of one of the proteins in your pathway? Is it a term that was reported enriched for that pathway? What sort of pathway analysis is this? Does the tool report over/under represented BP GO terms? In what? Does it return a list of pathways with over/under represented diff expressed genes? Presumably, the statistical test will only show something if it is significant, which will be hard with very small pathways, but please edit your question and give us some more details about the analysis. $\endgroup$ – terdon Jul 12 '18 at 11:15
  • $\begingroup$ statistically. the tool only returns fisher exact two tail test value and I already filtered my list of pathways according to that. the tool also returns pathways as go terms and the example go term is one of the significantly enriched $\endgroup$ – Fırat Uyulur Jul 12 '18 at 14:13
  • $\begingroup$ There's nothing special about the term. In order to help you, we would need to see the relevant pathway and the numbers involved. As I said before, please edit your question and give us more details. $\endgroup$ – terdon Jul 12 '18 at 14:18
  • $\begingroup$ Are you talking about pathway or about genes related to a GO term? It is not the same as pathway, although if you give us your definition of pathway it might be that it fits... If you don't take into account the structure of GO terms you'll get many false positives. Also, could you edit your question and include the data in text format? It will be easier to search more information about the genes and GO terms.. Also did you correct by multiple comparisons (as an answer suggest)? $\endgroup$ – llrs Jul 13 '18 at 7:23
4
$\begingroup$

The problem here is that any one GO term with a small number of members is unlikely to include any differentially expressed genes by chance, but if you test a large number of GO terms, some of them definitley will. The solution is to correct the p-values for multiple comparisons. If you do this, then none of those GO terms with a single DE member will be significant. The larger GO terms with multiple DE members in the chart you show probably won't survive correction either. Hopefully you have some that are more significant.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.