Retraining Augustus with BUSCO

Question

[Ric]

Hi, I tried to follow this instruction on how to retrain Augustus with BUSCO. I ran BUSCO:

run_BUSCO.py -i NbV1Ch07.fasta -o augustus_busco -l /assembly/busco-embryophyta_odb9/ -m geno --cpu 1 --long -sp coyote_tobacco --augustus_parameters='--progress=true'

and it created a folder /augustus/run_augustus_busco/augustus_output/retraining_parameters with the following content:

BUSCO_augustus_busco_3750827137_weightmatrix.txt
BUSCO_augustus_busco_3750827137_parameters.cfg.orig1
BUSCO_augustus_busco_3750827137_parameters.cfg
BUSCO_augustus_busco_3750827137_metapars.utr.cfg
BUSCO_augustus_busco_3750827137_metapars.cgp.cfg
BUSCO_augustus_busco_3750827137_metapars.cfg
BUSCO_augustus_busco_3750827137_intron_probs.pbl
BUSCO_augustus_busco_3750827137_igenic_probs.pbl
BUSCO_augustus_busco_3750827137_exon_probs.pbl

Which file should I use to run Augustus trained with BUSCO and how the command looks like?

Thank you in advance,

Answer 1

[jean.elbers]

You copy the files from /augustus/run_augustus_busco/augustus_output/retraining_parameters to your augustus installation directory under /augustus-install-directory/config/species/BUSCO_augustus_busco_3750827137/

usage:
augustus [parameters] --species=BUSCO_augustus_busco_3750827137 queryfilename

or if you want to output to be redirected to a file (see also outfile parameter):
augustus [parameters] --species=BUSCO_augustus_busco_3750827137 queryfilename > output.gff

[Ric]

Thank you. I ran BUSCO with -sp coyote_tobacco option. Should I have removed it and should I used a hard masked repeat assembly file for retraining Augustus?

[jean.elbers]

You have to run BUSCO with some species in the -sp setting. You should use already available augustus species in the -sp setting that is likely close to your assembly. You can't remove -sp and replace it with a hard masked repeat assembly file for retraining Augustus. The assembly you chose to run BUSCO with is NbV1Ch07.fasta. I don't think a hard masked repeated version of NbV1Ch07.fasta would improve results, but I have only ever run BUSCO with unrepeat masked assemblies.

[Ric]

Thank you. Do you know how to add UTR training?

[jean.elbers]

Sorry, I don't think you can train Augustus for UTRs with BUSCO.