Help using BUSCO to assess quality of transcriptome

Question

[RNAfan]

Hello,

I am trying to use BUSCO to assess the quality of a transcriptome assembled with Trinity. My run fails after ~2h running, when the HMMER phase begins. I couldn't find any support options for BUSCO.

Is there any way to get support for BUSCO? I couldn't find anything on line or on the documentation.

This is my command and programs versions:

module load Python/3.2.1
module load BLAST+/2.2.31
module load HMMER/3.1b2
module load EMBOSS/6.5.7

python3 ~/MyPrograms/BUSCO_v1.1b1/BUSCO_v1.1b1.py -in ../Trinity-GG.fasta -o GG \
-l ~/MyPrograms/BUSCO_v1.1b1/vertebrata -m trans -c 4 -f

After ~2 hours running, this is the error I get:

*** Running HMMER to confirm transcript orthology ***
Traceback (most recent call last):
  File "/.../MyPrograms/BUSCO_v1.1b1/BUSCO_v1.1b1.py", line 594, in <module>
    name=i[:-7];group=transdic[name]
KeyError: 'TRINITY_'

Any help would be appreciated

Thanks

[Rohit]

Just curious, but did you try with the version 1.1 instead of beta? I had problems with busco but it was usually due to using latest Augustus release (species model config was missing) instead of an older one. Also, probably it has to do with the sequence names in your trinity assembly, did you check that?

Answer 1

[RNAfan]

Hi

Thank you both for your suggestions. I was able to solve this without using docker.

I tried version 1.1 but I got the same error.

The newer versions of Trinity dropped the troublesome pipe (|) from their headers, so most compatibility issues are now solved. However, the problem was indeed in the header: it won't accept Trinity's header for genome-guided assemblies:

TRINITY_GG_XXXXX

Deleting _GG worked for me:

sed s/TRINITY_GG_/TRINITY_/ Trinity-GG.fasta > Trinity-GGbusco.fasta

I had no problems with running BUSCO with DeNovo assemblies, which have the header TRINITY_DNXXXXXXXXX

Thanks

Answer 2

[Nestor Wendt]

Hello, I ran into this problem yesterday but could not figure it out. However, when I tried running BUSCO from this Docker container, it worked.

Once you have Docker installed, just run something like that:

docker run -v /path/to/workdir/:/input -v /path/to/workdir/:/results cfriedline/busco python3 /busco/BUSCO_v1.1b1/BUSCO_v1.1b1.py -o example -in /input/example.fasta -l /busco/dataset -m trans -c 4 -e 1e-05

Change the parameters accordingly.