Can't select 0.0 nor 0.00 in model_selection_value in AbinitioTraining.nf

[LucileSol]

Hi! I wanted to select AED only equal to 0 and got the following error:

Error executing process > 'abinitio_training:blast_makeblastdb (codingGeneFeatures.filter.longest_cds.complete.good_distance_proteins type: null)'

Caused by:
  Process `abinitio_training:blast_makeblastdb (codingGeneFeatures.filter.longest_cds.complete.good_distance_proteins type: null)` terminated with an error exit status (1)

Command executed:

  makeblastdb -in codingGeneFeatures.filter.longest_cds.complete.good_distance_proteins.fasta -dbtype prot

Command exit status:
  1

Command output:

  Building a new DB, current time: 04/24/2020 09:39:23
  New DB name:   /scratch/nxf.EEhdsta03j/codingGeneFeatures.filter.longest_cds.complete.good_distance_proteins.fasta
  New DB title:  codingGeneFeatures.filter.longest_cds.complete.good_distance_proteins.fasta
  Sequence type: Protein
  Keep MBits: T
  Maximum file size: 1000000000B

Command error:
  BLAST options error: File codingGeneFeatures.filter.longest_cds.complete.good_distance_proteins.fasta is empty

Work dir:
  /projects/annotation/agrotis_infusa/abinitio/work/f6/2ec5d6f0ce7146a67310e30a9cde28

Tip: view the complete command output by changing to the process work dir and entering the command `cat .command.out`

I have AED being only 0.00 and it is working with if I choose 0.01 as parameter

[Juke34]

Probably a bug in AGAT

[mahesh-panchal]

Yeah, it looks like an empty fasta was written by the previous step. You need to trace back through the outputs in the work folders to find which stage is causing the erroneous output.

[LucileSol]

04/24/2020 at 10h01m36s
usage: /sw/anaconda/2019.10/envs/agat/bin/agat_sp_filter_feature_by_attribute_value.pl --gff mrna.gff --value 0.00 -a _AED -t >= -o codingGeneFeatures.filter.gff
We will discard all features that have the attribute _AED with the value >= 0.00.
550877 features removed 

it removed everything because of the = should we only use > in the pipeline?

[Juke34]

It's a possibility

[mahesh-panchal]

If you both agree on it, then just edit the master code and reference this issue. It's one deletion.

[LucileSol]

either we use only > and we know that we will not include the value that we put in the parameter or we keep >= and make a special case for 0.00

[Juke34]

The first sounds better

[LucileSol]

sorry I made a mistake we use only > and we know that we are including the value we put in the parameter (it is when we do >= that we exclude it). I am fine with it

[Juke34]

So, is there any error in AGAT to solve?

[LucileSol]

no just the pipeline, to remove the =

[mahesh-panchal]

Sorted