I understand the over-representation of GO term in FDR test against reference gene sets but the under-representation seems confusing. My test set is infection induced differentially expressed host genes (DEG).
Is the under-representation convey the message "X function is will not occur as this shows under-representated GO categories" of DEG [up and downregulated]?
A GO annotation is … "a statement that a gene product";
1- Has a particular molecular function .. 'OR'
2- Is involved in a particular biological process .. 'OR'
3- Is located within a certain cellular component ,,, as determined by a particular method .. OR as described in a particular reference.
When you have a set of differentially expressed (DE) genes related to a specific condition, you usually want to infer their biological meaning and whether they are related in some way. If there are more DE genes for a particular category, then this category is "enriched" and the associated P-value is called "over-represented", while if there are fewer DE genes for a particular category, then this category is called "depleted" and "under-represented". This is usually denoted by a ‘+’ or ‘-‘ symbol (i.e., direction) for the over- and under-represented categories, respectively.
Most popular type of GO analysis determines which GO terms are more often associated with a specified list of genes/proteins compared with a control list (you have to decide which reference/control list to use as your background; built-in in the tool or your own list).
As for the under-represented GO terms (ex: certain signaling pathway), they might be as important as the over-represented ones if they are ‘strongly’ under-represented (with a highly significant P-value) as this means that normal functioning of that pathway is necessary for the given condition.
Gene set enrichment analysis (GSEA) can help in that regard as it doesn’t require a priori list of selected genes. Instead, it analyzes your whole list of transcripts (genes) –both DE and nonDE- against thousands of curated pathways and GO terms. The result will give you an idea on which gene sets (terms or pathways) are associated with your treatment (‘+’ or ‘-‘); usually using an enrichment score (ES).
Hope this helps a bit.
R
Useful source: http://www.biotnet.org/sites/biotnet.org/files/documents/158/goa_tutorial_april2013.pptx
From the way you're raising, it has a limited biological sense. I think that the under-representation in this case means the contrapositive affirmation: "this function is not modified for the treatment". In rare cases, it could be consider as: "the normal function it is necesary at a given condition". But if you were talking only about upregulated or downregulated, then the under-representation of a biological process could be found more meaningful.
Dear Asthmit,
I think it depends on the genes and their role in the particular test SAMPLE. kEGG metabolic pathways can help to correlate the under-expression of genes and the reasons depending on the the gene product function.
A GO annotation is … "a statement that a gene product";
1- Has a particular molecular function .. 'OR'
2- Is involved in a particular biological process .. 'OR'
3- Is located within a certain cellular component ,,, as determined by a particular method .. OR as described in a particular reference.
When you have a set of differentially expressed (DE) genes related to a specific condition, you usually want to infer their biological meaning and whether they are related in some way. If there are more DE genes for a particular category, then this category is "enriched" and the associated P-value is called "over-represented", while if there are fewer DE genes for a particular category, then this category is called "depleted" and "under-represented". This is usually denoted by a ‘+’ or ‘-‘ symbol (i.e., direction) for the over- and under-represented categories, respectively.
Most popular type of GO analysis determines which GO terms are more often associated with a specified list of genes/proteins compared with a control list (you have to decide which reference/control list to use as your background; built-in in the tool or your own list).
As for the under-represented GO terms (ex: certain signaling pathway), they might be as important as the over-represented ones if they are ‘strongly’ under-represented (with a highly significant P-value) as this means that normal functioning of that pathway is necessary for the given condition.
Gene set enrichment analysis (GSEA) can help in that regard as it doesn’t require a priori list of selected genes. Instead, it analyzes your whole list of transcripts (genes) –both DE and nonDE- against thousands of curated pathways and GO terms. The result will give you an idea on which gene sets (terms or pathways) are associated with your treatment (‘+’ or ‘-‘); usually using an enrichment score (ES).
Hope this helps a bit.
R
Useful source: http://www.biotnet.org/sites/biotnet.org/files/documents/158/goa_tutorial_april2013.pptx
Dear Ramy, thanks, this is very helpful. I could not understand this paragraph "
"As for the under-represented GO terms (ex: certain signaling pathway), they might be as important as the over-represented ones if they are ‘strongly’ under-represented (with a highly significant P-value) as this means that normal functioning of that pathway is necessary for the given condition."
Could you please explain it little more. I have some GO categories underrepresented during bacterial infection to host.
"
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology