Errors in file embl

Question

[kamel]

Hi biostar,

Maybe basic questions but I can not solve this problem. I wanted to validate an embl file. After the validation I find two errors (below)

ERROR: Intron is usually expected to be at least 10 nt long ERROR: Abutting features can not be adjacent between neighboring exons.

Have an idea about how I could fix one or both errors.

Thank you

[lieven.sterck]

you can find some pointers in this blog post already : http://avrilomics.blogspot.com/2018/08/converting-gff3-file-to-embl-to-submit.html

Answer 1

[Juke34]

For the intron size problem, you can fix that by adding the /pseudo attribute. Have a look here to understand the meaning of this attribute.
If you want I have a script that do it automatically.
It is called gff3_sp_flag_short_intron.pl and you can find it in the GAAS repository.
It is called agat_sp_flag_short_introns.pl and is part of the of GFF toolkit AGAT.

For this Abutting features error, if the two features are from different genes then it is a bug from their validator. Have a look here. They advise to remove the exons from the flat file to avoid that type of problem. Otherwise you have to modify location manually.

[kamel]

Thank you @jike-34 for your reply. I installed all perl dependencies of GAAS to use your scripts gff3_sp_flag_short_intron.pl. I ran it but it also asks the module "BILS :: CheckModule" but mcpan it can not find this module

[Juke34]

It just mean that you haven't completed the installation procedure.
Everything starting by BILS is related to a personal library that is within the GAAS repo. To access them (because it is not the only one needed) 2 choices:

cd GAAS
make install
source profiles/activate_env

or add in your ~/.bashrc or ~/.profile file (provide the correct path):

export PERL5LIB=$PERL5LIB:/pathTo/GAAS/annotation export
PATH=${PATH}:/pathTo/GAAS/annotation/Tools/bin

[kamel]

Thanks @Juke-34 for your help. It is well installed. I have a quick question, Do you think that the argument --translate of EMBLmyGFF3 is mandatory for submission?

[Juke34]

Did you run gff3_sp_flag_short_intron.pl without problem this time ?
No --translate it is not mandatory for submission.

[kamel]

Thank you for your reply. Yes I ran gff3_sp_flag_short_intron.pl without any problem. Except sometimes it put twice / pseudo in the same CDS which is not recommended for a valid file, but it works very well. Thanks again.

[Juke34]

I forgot about this duplication of attributes within the EMBL output. Did you find a way to fix it ? Because the validator does not like them... I have fixed that within EMBLmyGFF3 but I didn't push the fix yet...

[kamel]

yes the validator found this as an error, unfortunately I delete them manually

[Juke34]

Fix pushed in EMBLmyGFF3 version 1.2.5