trans-spliced mRNA in chloroplast genome
Closed this issue · 2 comments
Thank you for developing this wonderful tool. I usually use liftoff to transfer annotations for eukaryotic nuclear genomes and always get nice results.
I'd like to report an issue while lifting over a trans-spliced mRNA in a chloroplast genome.
Following is the gene structure in problem in the reference genome.
Chloroplast feature gene 1 66800 . + . ID=MpKit2_Cp001;locus_tag=MpKit2_Cp001
Chloroplast feature mRNA 1 66800 . + . ID=MpKit2_Cp001.1;Parent=MpKit2_Cp001;codon_start=1;gene=rps12;locus_tag=MpKit2_Cp001;note=30S ribosomal protein S12%3B rps12 CDS%3B trans splicing of rps12 intron 1;product=ribosomal protein S12;protein_id=BBD75102.1;trans_splicing=;transl_table=11;translation=MPTIQQLIRNKRQPIENRTKSPALKGCPQRRGVCTRVYTTTPKKPNSALRKIARVRLTSGFEITAYIPGIGHNLQEHSVVLVRGGRVKDLPGVRYHIIRGTLDAVGVKDRQQGRSKYGVKKSK
Chloroplast feature intron 65901 66686 . - . ID=MpKit2_Cp001.1.intron1;Parent=MpKit2_Cp001.1
Chloroplast feature CDS 66687 66800 . - 0 ID=MpKit2_Cp001.1.cds1;Parent=MpKit2_Cp001.1
Chloroplast feature intron 1 92 . + . ID=MpKit2_Cp001.1.intron2;Parent=MpKit2_Cp001.1
Chloroplast feature CDS 93 324 . + 0 ID=MpKit2_Cp001.1.cds2;Parent=MpKit2_Cp001.1
Chloroplast feature intron 325 828 . + . ID=MpKit2_Cp001.1.intron3;Parent=MpKit2_Cp001.1
Chloroplast feature CDS 829 854 . + 0 ID=MpKit2_Cp001.1.cds3;Parent=MpKit2_Cp001.1
Here, the first exon is encoded in the - strand, while other 2 exons are encoded in the + strand.
When this gene was lifted-over using liftoff, all the exons were encoded in the + strand, and the exons were ordered by the position in the genome, which was different from the original order in the reference.
CP Liftoff gene 1 66802 . + . ID=MpKit2_Cp001;locus_tag=MpKit2_Cp001;coverage=1.0;sequence_ID=1.0;valid_ORFs=0;extra_copy_number=0;copy_num_ID=MpKit2_Cp001_0
CP Liftoff mRNA 1 66802 . + . ID=MpKit2_Cp001.1;Parent=MpKit2_Cp001;codon_start=1;gene=rps12;locus_tag=MpKit2_Cp001;note=30S ribosomal protein S12; rps12 CDS; trans splicing of rps12 intron 1;product=ribosomal protein S12;protein_id=BBD75102.1;trans_splicing=;transl_table=11;translation=MPTIQQLIRNKRQPIENRTKSPALKGCPQRRGVCTRVYTTTPKKPNSALRKIARVRLTSGFEITAYIPGIGHNLQEHSVVLVRGGRVKDLPGVRYHIIRGTLDAVGVKDRQQGRSKYGVKKSK;matches_ref_protein=False;valid_ORF=False;missing_start_codon=True;extra_copy_number=0
CP Liftoff CDS 93 324 . + . ID=MpKit2_Cp001.1.cds2;Parent=MpKit2_Cp001.1;extra_copy_number=0
CP Liftoff CDS 829 854 . + . ID=MpKit2_Cp001.1.cds3;Parent=MpKit2_Cp001.1;extra_copy_number=0
CP Liftoff CDS 66689 66802 . + . ID=MpKit2_Cp001.1.cds1;Parent=MpKit2_Cp001.1;extra_copy_number=0
CP Liftoff intron 1 92 . + . ID=MpKit2_Cp001.1.intron2;Parent=MpKit2_Cp001.1;extra_copy_number=0
CP Liftoff intron 325 828 . + . ID=MpKit2_Cp001.1.intron3;Parent=MpKit2_Cp001.1;extra_copy_number=0
CP Liftoff intron 65903 66688 . + . ID=MpKit2_Cp001.1.intron1;Parent=MpKit2_Cp001.1;extra_copy_number=0
Although trans-spliced genes are not so common in nuclear genomes, they are frequently observed in organellar genomes. Hopefully, liftoff can handle trans-splicing in the future version.
I used liftoff ver. 1.6.1.
To reproduce this, the FASTA and GFF files available from here can be used.
(The sequence name in FASTA needs to be changed to 'Chloroplast')
Hi,
Currently liftoff is really not designed to handle trans-splicing. This issue is that because the 'gene' and 'mRNA' features are annotated on the + strand, liftoff expects the child features to be on the + strand too. One work around would be to remove the Parent= field from the introns and CDSs and the include 'intron' and 'CDS' in a text file with -features. This will make it so that each of these features is lifted over independently which will allow them to map to different strands.
Thank you for the suggestion. I tried the workaround, removing Parent field and specifying an "-f" option (Does "-features" mean "-f"?).
But it failed with an error like this.
File "path/to/liftoff/polish.py", line 121, in group_cds_by_tran
groups[cds.attributes["Parent"][0]].append(cds)
KeyError: 'Parent'
So removing "Parent" field does not seem to work.
Anyway, I was able to transfer the annotation by splitting gene and mRNA depending on which strand they are encoded.