training GeneMark for gene prediction

[ampsevilla]

Hi,

I'm want to annotate a big, complex eukaryotic genome from non-model using GeneMark 4.69 gene prediction.

I want to train GeneMark algorhytm using data from a close-related specie using its data (its genome sequence in FASTA format and its genes annotations in GFF3 and GFF format), but when I try it with using this command (or similars)

gmes_petap.pl --ET ref/INCAR_Mch_V1.gff --sequence ref/INCAR_Mch_V1.fa > prob/predicted.gff

I always get an error related to the format of sequence or so.

I have got lost about how to train GeneMark from FASTA and GFF3 files. Could someone help me?

Thank you so much.