Annotate A Genome With A Well Annotated Related Species

Question

[Nicolas Rosewick]

Hi,

What is the easiest way to annotate a genome that has no annotation. I've the genome and annotation of a very related species.

In final, I wan to have the same structure as refGene.txt and ensGene.txt from UCSC.

Here's an example from the RAT ensGene.txt :

585    ENSRNOT00000044270    chr1    +    5472    16844    5472    16844    8    5472,5523,8281,8991,10768,14683,15968,16733,    5485,5725,8567,9795,10993,14807,16732,16844,    0    ENSRNOG00000021490    incmpl    incmpl    0,1,2,0,0,0,1,0,
585    ENSRNOT00000049921    chr1    +    5525    16968    5525    16968    8    5525,8275,8991,10768,14677,15965,16733,16943,    5725,8567,9795,10993,14807,16732,16846,16968,    0    ENSRNOG00000021490    cmpl    incmpl    0,2,0,0,0,1,0,2,
585    ENSRNOT00000051735    chr1    +    5525    16968    5525    16968    7    5525,8275,8991,10768,15965,16733,16943,    5725,8567,9795,10997,16732,16846,16968,    0    ENSRNOG00000021490    cmpl    incmpl    0,2,0,0,1,0,2,

My strategies is to download all the gene (CDNA ?) from ensembl and ucsc from this related species and to map with gmap them on the unannotated genome. Is it a good way to do this ?

Thanks a lot,

N.

Answer 1

[Lhl]

yes, basically you can follow the similarity-searching-based (map proteome of relative species to your genome assembly) method. And you can also try some de novo gene prediction approaches. Then, it is good to combine the two different strategies (this can be done using some combiners, e,g,, EVM, glean, Maker ...).