I have to analyze using Gene ontology a series of genes, of which I have their Uniprot code. I have been able to put these codes in Gene Oncology, indicating the organism to which it belongs (Streptococcus pneumoniae). I don't know how to analyze the next data obtained, If, for example, it appears to me that in transport proteins .90 is expected and I obtain 6.70, what I have to comment is that after the treatment there is that increase? And what does # indicate?
The typical scenario is that you have some genes which changed after treatment - let's say you did an experiment and found that 115 genes from Streptococcus changed, and these are what you bring for analysis with the resulting table above. The question is: what Ontology terms are enriched in this set of 115 genes? To answer this, we have to know the expected frequency of finding any GO term in a random sample of 115 Streptococcus genes by chance. We can determine this from the first column of your table: there are 18 genes annotated to Protein Transport, and Streptococcus has about 2200 genes, so if we select 115 genes randomly from a cookie jar, we would expect to find: 115 * 18/2200 of them to be associated to Protein Transport just by chance. Thus 0.9 is the expected number of genes in a random draw of 115 genes. But you didn't do a random draw - you did an experiment to get those 115 genes, and the number of genes associated to Protein Transport in your 115 gene set is 6. This is your observed number. Thus you have 6/0.9 = 6.7 more than you expect from random chance (observed/expected).
So, yes, you have an increase of that gene type after treatment. And "#" reflects the number of genes of that term type observed in your particular gene set. The resulting p-value could come from a Fisher test, or the hypergeometric distribution.
There is also a thread here with a few additional details that may aid your understanding. Mostly, it highlights the importance of the background geneset.
Thank you very much for your time and for your explanation, it has helped me a lot!
There is also a thread here with a few additional details that may aid your understanding. Mostly, it highlights the importance of the background geneset.
Thank you for the link, it's very useful!