Maker2 CDS outside of gene in GFF3

Question

[plattsad]

Hi, I've been annotating a genome with Maker (2.31.10) and noticed that amongst the 30k gene models annotated, there are 9 where one or more CDSs lie outside of the coordinates defined by the gene line in the gff3. This seems odd to me, has anyone encountered this and would it be a glitch or a valid model? I'm getting this when I parse the final gff3 using either gaas_maker_merge_outputs_from_datastore.pl or gff3_merge.

There's an example below. The gene model 20.71 runs from 2,044,040 to 2,046,575, but CDS 1 of mRNA-1 runs from 2,042,847 to 2,043,849 which is at its closest ~200bp upstream of the gene's coordinates.

1   maker   gene    2044040 2046575 .   +   .   ID=maker-1-augustus-gene-20.71;Name=maker-1-augustus-gene-20.71
1   maker   mRNA    2042847 2046575 .   +   .   ID=maker-1-augustus-gene-20.71-mRNA-1;Parent=maker-1-augustus-gene-20.71;Name=maker-1-augustus-gene-20.71-mRNA-1;_AED=0.33;_eAED=0.35;_QI=0|0|0.16|0.83|1|1|6|0|571
1   maker   mRNA    2044040 2046575 .   +   .   ID=maker-1-augustus-gene-20.71-mRNA-2;Parent=maker-1-augustus-gene-20.71;Name=maker-1-augustus-gene-20.71-mRNA-2;_AED=0.43;_eAED=0.49;_QI=0|0|0.2|0.8|1|1|5|0|242
1   maker   exon    2042847 2043849 .   +   .   ID=maker-1-augustus-gene-20.71-mRNA-1:exon:958;Parent=maker-1-augustus-gene-20.71-mRNA-1
1   maker   exon    2044056 2044179 .   +   .   ID=maker-1-augustus-gene-20.71-mRNA-1:exon:959;Parent=maker-1-augustus-gene-20.71-mRNA-1
1   maker   exon    2044944 2045008 .   +   .   ID=maker-1-augustus-gene-20.71-mRNA-1:exon:960;Parent=maker-1-augustus-gene-20.71-mRNA-1,maker-1-augustus-gene-20.71-mRNA-2
1   maker   exon    2045530 2045817 .   +   .   ID=maker-1-augustus-gene-20.71-mRNA-1:exon:961;Parent=maker-1-augustus-gene-20.71-mRNA-1,maker-1-augustus-gene-20.71-mRNA-2
1   maker   exon    2046186 2046266 .   +   .   ID=maker-1-augustus-gene-20.71-mRNA-1:exon:962;Parent=maker-1-augustus-gene-20.71-mRNA-1,maker-1-augustus-gene-20.71-mRNA-2
1   maker   exon    2046421 2046575 .   +   .   ID=maker-1-augustus-gene-20.71-mRNA-1:exon:963;Parent=maker-1-augustus-gene-20.71-mRNA-1,maker-1-augustus-gene-20.71-mRNA-2
1   maker   exon    2044040 2044179 .   +   .   ID=maker-1-augustus-gene-20.71-mRNA-2:exon:964;Parent=maker-1-augustus-gene-20.71-mRNA-2
1   maker   CDS 2042847 2043849 .   +   0   ID=maker-1-augustus-gene-20.71-mRNA-1:cds;Parent=maker-1-augustus-gene-20.71-mRNA-1

TIA for suggestions.

Answer 1

[Juke34]

Sounds like a bug. You should report the problem to the developer. It is maybe fixed in MAKER version 3. You can run all those cases by running agat_convert_sp_gxf2gxf.pl from AGAT:

agat_convert_sp_gxf2gxf.pl --gff input.gff -o output_clean.gff