How to show biological processes associated with a small list of genes?

Question

[Diwan]

Hello,

I have a small list of 20 genes and I want to show (or make a statement about) the biological process associated with them. GO/KEGG enrichment analysis doesn't work as the gene list is small. Any suggestions?

How do you show the biological process associated with a small list of genes?

Thanks for your help,
Diwan

Answer 1

[Antonio R. Franco]

I don't fully understand.

The only way I see feasible to make a statement about the biological processes associated to a group of proteins, being a small group or a big group, is by using the information of the GO annotations. If the group is big, you need the help of an enrichment analysis or Fischer test

But if you work with a small group or proteins as is your case, then it has to be ease for you to figure out how similar are their biological function without having to run a comparison or a complex task

If you see that all of the GO are similar, you only get a conclusion.

If they are different, you get another different conclusion

That being said it is important to say that you need to know what are you comparing.

GO annotations navigate among different levels. Then, you can have many proteins associated to a same GO level (e.g, proteins corresponding to the general nitrogen metabolism (Level2) ) when they are actually different proteins that are part of this level, such as nitrogenase, nitrate reductase, nitrite reductase, and so on (that is levels 3 or 4)

Answer 2

[Jean-Karim Heriche]

Any experienced biologist should be able to tell you what processes are represented in a list of 20 genes. Just find a friendly one in your neighborhood and ask them. You don't necessarily need GO for everything.

Answer 3

[Renesh]

The one way to say by statistically is to compare the biological processes among 20 genes only.

• Get the GO annotation for those 20 genes
• cluster the biological process for each individual Go term under biological process
• Apply the simple statistics (t-test or any another to find which is statistically different from another)

I know this will not be robust statistically as your simple size is small. But it will give an overall idea which biological processes are highly represented in your gene list.