Eukaryotic Genome Annotation for a non-model organism

Question

[flomascali]

Hi, I have a genome assembly of 1.2 Gb for a non-model fish and I want to annotate it. I 've already done a de novo prediction with AUGUSTUS and then a blastp search with the Zebrafish's proteins. I want to know, wich other strategies do you recommend me? For example, does it make any sense to do a tblastn search between my contigs and the Zebrafish proteins? And in that case, wich parameters do you recommend? Thanks!

[Mehmet]

What is your aim? What kind of additional data you have and/or you can have? Depending on your aim you can determine which analyses are needed.

Answer 1

[Mensur Dlakic]

You may want to take a look at MAKER, especially if you have access to ESTs from the same or related organism. There is a tutorial that will give you an idea of what is needed in terms of resources and what output to expect.

Answer 2

[colindaven]

Maker is good, as suggested by others. It can integrate gene predictions by Augustus etc and importantly, use their evidence, eg, proteins and transcripts.

You can generate transcript sets by yourself if you have RNA-seq (or publicly available RNA-seq) - try Trinity - and align these to the genome with GMAP (with gff3 output).

You'll need to get used to easy standard formats like GFF3, GTF and BED, and make data availabe to your coworkers using a genome browser (I like JBrowse, IGV is another option).

Versioning is also very important as time goes on, so version the assembly, the gene sets, transcript sets etc.