I am using MG-RAST to analyse metagenome datasets and have used the Subsystems, KO, NOG, and COG databases for annotation of the functional genes. However, I don't know enough about the different databases so to choose which would be the most reliable library for my dataset. I have environmental samples and am determining microbial metabolic potential in terms of hydrocarbon degradation.
I can't say what could be the best choice for your samples but I can tell about the difference. KO is hosted by KEGG and is manually curated and assigned molecular functions based on KEGG pathways. COG was initially generated by protein sequences of microbial genomes with known and predicted proteins. This structure was used to create other databases like KO and NOG (eggNOG).
NOG database can created by the machine learning non supervised approach.
All these databases are fairly up to data and I don't think there could a lot of differences among them. COGs was especially for microbes and KOG and NOG also contain the eukaryotic functional groups.
Just an addition to Abhijeet Singh , the subsystems technology used by mgRast follows a logic that is somewhat different from other databases, such as NOG, KO and eggNOG, so they are not comparables. You should chose one and stick with it.
- Badges
- Science topic
- Similar topics
- Chemistry
- Biochemistry