empty attributes are appended to feature ID

[Rik-149]

agat_convert_sp_gxf2gxf.pl appends empty attributes to the feature ID. I noticed this when running the script on an EviAnn annotation, this pipeline adds a class attribute to the mRNA features which is sometimes "Class==", my guess is that the parser gets confused here and assumes the Class attribute is empty.

What I cannot explain is that this is then added to the feature ID. XLOC_011812-mRNA-1 becomes XLOC_011812-mRNA-1,Class

I am using the agat 1.5.1 container from quay.io

agat_config.yaml

Original file:

SL4.0ch08       EviAnn  mRNA    54575899        54580638        .       +       .       ID=XLOC_011812-mRNA-1;Parent=XLOC_011812;EvidenceProteinID=XLOC_017317U-mRNA-1;EvidenceTranscriptID=MSTRG_00045729:4:54.736046;StartCodon=ATG;StopCodon=TGA;Class==;Evidence=transcript_only
SL4.0ch08       EviAnn  exon    54575899        54576107        .       +       .       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577093        54577135        .       +       .       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577215        54577307        .       +       .       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577402        54577464        .       +       .       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577754        54577849        .       +       .       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577948        54578025        .       +       .       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54580334        54580638        .       +       .       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54575974        54576107        .       +       0       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577093        54577135        .       +       1       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577215        54577307        .       +       0       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577402        54577464        .       +       0       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577754        54577849        .       +       0       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577948        54578025        .       +       0       Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54580334        54580420        .       +       0       Parent=XLOC_011812-mRNA-1

after agat_convert_sp_gxf2gxf.pl

SL4.0ch08       EviAnn  mRNA    54575899        54580638        .       +       .       ID=XLOC_011812-mRNA-1,Class;Parent=XLOC_011812;evidence=transcript_only;evidenceProteinID=XLOC_017317U-mRNA-1;evidenceTranscriptID=MSTRG_00045729:4:54.736046;startCodon=ATG;stopCodon=TGA
SL4.0ch08       EviAnn  exon    54575899        54576107        .       +       .       ID=agat-exon-94261;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54575974        54576107        .       +       0       ID=agat-cds-82023;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577093        54577135        .       +       1       ID=agat-cds-82024;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577093        54577135        .       +       .       ID=agat-exon-94262;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577215        54577307        .       +       0       ID=agat-cds-82025;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577215        54577307        .       +       .       ID=agat-exon-94263;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577402        54577464        .       +       0       ID=agat-cds-82026;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577402        54577464        .       +       .       ID=agat-exon-94264;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577754        54577849        .       +       0       ID=agat-cds-82027;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577754        54577849        .       +       .       ID=agat-exon-94265;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54577948        54578025        .       +       0       ID=agat-cds-82028;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54577948        54578025        .       +       .       ID=agat-exon-94266;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  CDS     54580334        54580420        .       +       0       ID=agat-cds-82029;Parent=XLOC_011812-mRNA-1
SL4.0ch08       EviAnn  exon    54580334        54580638        .       +       .       ID=agat-exon-94267;Parent=XLOC_011812-mRNA-1

Whenever the Class attribute has a different value "Class=j", this behavior does not occur and the feature ID is not changed.

[Juke34]

I will add a patch but in this case (Class==) the Class attribute will disappear... indeed, value = is forbidden and must be escaped as %3D. It is mentioned in GFF3 specification https://github.com/The-Sequence-Ontology/Specifications/blob/master/gff3.md
So a warning will be thrown i.e. "No value for tag Class in GFF3 attribute"