Eukaryotic gene prediction using GeneMark-ES

Question

[lakhujanivijay]

I am using GeneMark-ES Suite version 4.33 for Eukaryotic gene prediction for plant genome and I am kind of stuck

command

perl gmes_petap/gmes_petap.pl --evidence protein.fa  --cores 40 --sequence genome_assembly.fa --ET transcripts.gff

---

input files

protein.fa = a multi-fasta file having amino acid sequences from a closely related plant

genome_assembly.fa = genome assembly multi-fasta file having scaffold sequences for which I want to predict the genes

transcripts.gff = gff file for transcript sequences

---

error message

error, unexpected format found on line: >prot.1

error on call: /gmes_petap/reformat_gff.pl --out data/evidence.gff  --trace info/dna.trace  --in protein.fa  --quiet

---

I think I am providing a wrong file in the --evidence parameter as shown below

--ET           [filename]; to run training with introns coordinates from RNA-Seq read alignments (GFF format)
--evidence     [filename]; to use in prediction external evidence (RNA or protein) mapped to genome

What could (RNA or protein) mapped to genome possibly mean? Any ideas?

Answer 1

[lieven.sterck]

Not an expert in using GeneMark myself but I'm guessing it assumes some kind of alignment file (in gff format) of the proteins to the genome, and thus not a protein fasta file (which you apparently are providing)

[lakhujanivijay]

An alignment file in gff format sounds alien to me.

[lieven.sterck]

well, I meant alignment as in 'HSP coordinates of aligned proteins to the genome' ( obtained by using eg. blast (not recommended), GenomeThreader, GeneWise ... ). apologies for the brevity .