Presently, there are many methods to calculate functional similarity (semantic similarity) between the genes. But there are three domains associated with any gene i.e., Molecular Function (MF), Biological Process (BP) and Cellular Component (CC).
When performing functional similarity what domain has to be considered? Why?
I would say MF or BP. CC just tells you if it's cytoplasm, nucleus, etc. and while that is important, looking at the function and process is going to tell you it's location. I.e. if process says transporter, then you know it's likely in the membrane. Also DAVID analytical tool helps a lot and will group similar genes together based on ontology.
I hope that helps!
The BP is usually the first goto. It has a lot of terms in comparison to CC which results in more fine-grained comparison. Also most terms are humanly readable. The MF terms are quite technical/chemical thus harder to compare. But still can provide useful semantic knowlege.
The CC terms can be informative if you want to spatially divide your genes inside the cell. E. g. two genes can have identical or similar MF terms, but are spatially far apart like nucleolus and mitochondrion.
You should also consider other ontologies, not just GO, for semantic similarity.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics