I have a de-novo assembled Transcriptome for four different time points. What is the best way annotate it, right now I am using Trinotate. What percentage of transcripts should I expect to be present in the data. This organism has no annotated genome, so what e-value criteria should I used,
The assembled transcriptome contain partial gene/isoforms. I will suggest you first create a list if unigenes by sequence clustering. It will give you the idea about the genes
Then you can use the blast2go for annotation and classification.
Hello, as suggested Blast2Go would be an option. However keep in mind, that B2G annotations are not all the time very accurate, in that case you need to manually annotate (this procedure should be implemented where you think its necessary depending on your data, automated annotations are not always very accurate). Regarding the e-value: try with 1e-3 and 1e-6. Compare them, if all your seq. are annotated with 1e-6 then take that. Is there any genome of other closely related organism(s)?
Hi,
I personally prefer Trinotate over Blast2Go. What I don't like about B2G is that you have to generate the blast results in an XML file, while Trinotate contains all the information in a sql-lite database. Said that, you only have the annotation for their Swissprot database. However, the latest version of Trinotate allows you to add any custom blast search against any database, so you can blast your transcripts against any database and add that information. I prefer use Swissprot since it doesn't make any sense to annotate something as hypotetical protein.
Good luck.
- Badges
- Science topic