Using miniprot projections as evidence class in evidencemodeler

[yuzhenpeng]

Hello, @lh3

Thank you for developing and maintaining miniprot.

I created miniprot projections for one of my newly sequenced assemblies, which I wish to use in EVM for genome annotation. So I was wondering if you would be willing to share a script to convert miniprot output into a GFF3 file that EVM accepts? I thought I might ask if you already have a procedure developed for it.

Thanks in advance, Zhenpeng

[lh3]

Miniprot outputs GFF3. What does EVM require?

[CongLiu37]

Hello,

I have the same question. EvidenceModeler wants protein alignments in gff3. Example file can be found here https://evidencemodeler.github.io/

It would be great if miniprot can output both protein aligmnets and full-length gene structures with correct splice sites (gene, mRNA, exon, CDS, three_prime_cis_splice_site, five_prime_cis_splice_site) in two separate gff3 files. GenomeThreader can generate such files but need to run twice with different settings (-intermedia/-skipalignments) and is slow. Protein alignments can be integrated with other evidence by EvidenceModeler and gene structures can be used as training set for predictors like AUGUSTUS.

Many thanks!

Sincerely,

Cong

[lh3]

The evm example doesn't have the three_prime_cis_splice_site or the five_prime_cis_splice_site features. The only difference from the miniprot gff seems the gene and exon features. Could you write a script to add these two features? If you can confirm evm requires those two features, I can add them.

[CongLiu37]

Hello,

The protein alignment gff only has "match" feature and evm requires ID and Target in 9th field. It looks like this:

Contig1 nap-nr_minus_rice.fasta nucleotide_to_protein_match     8392    8470    50.00   -       .       ID=match.nap.nr_minus_rice.fasta.37;Target=RF|YP_440341.1|83716234|NC_007650 196 222
Contig1 nap-nr_minus_rice.fasta nucleotide_to_protein_match     7650    7786    26.09   -       .       ID=match.nap.nr_minus_rice.fasta.37;Target=RF|YP_440341.1|83716234|NC_007650 222 268

Contig1 nap-nr_minus_rice.fasta nucleotide_to_protein_match     8386    8509    26.83   -       .       ID=match.nap.nr_minus_rice.fasta.38;Target=RF|YP_099363.1|53713371|NC_006347 1 42
Contig1 nap-nr_minus_rice.fasta nucleotide_to_protein_match     7635    7786    24.00   -       .       ID=match.nap.nr_minus_rice.fasta.38;Target=RF|YP_099363.1|53713371|NC_006347 42 92

It is the second evm example for alignments instead of the first one for gene prediction gff with gene/exon features.

So it would be great if miniprot can output two gff files: one contains alignments and looks like what I pasted above, and another one contains gene structures with gene, exon, mRNA, exon, CDS features.

Many thanks!

Sincerely,

Cong

[lh3]

You can write a script to generate the second gff. This is highly evm specific and seems redundant. I don't add features for just one tool.