Hi Biobakery community!
I recently upgraded to HUMAnN v3.0 after having used HUMAnN2, but don’t seem to be able to run the demo because of a recurring error.
This text will be hidden
I installed it via conda in a contained environment, as suggested here: humann3 – The Huttenhower Lab
# create env
conda create --name biobakery3 python=3.7
# activate env
conda activate biobakery3
# update channels
conda config --add channels defaults
conda config --add channels bioconda
conda config --add channels conda-forge
conda config --add channels biobakery
# install biobakery and metaphlan
conda install -c biobakery human
conda install -c bioconda metaphlan
# test installation
humann_test
All of the tests seemed to be fine. I then proceeded to run the demo:
humann -i demo.fastq -o demo_outputs
The first time I ran it, it took 4 hours, after which the job was killed on the cluster I was using.
After that, it returned the following errors:
CRITICAL ERROR: Error executing: /data/home/btx143/.conda/envs/biobakery3/bin/metaphlan /data/home/btx143/shotgun2/demo/data/demo.fastq -t rel_ab -o /data/home/btx143/shotgun2/demo/outputs/demo_output/demo_humann_temp/demo_metaphlan_bugs_list.tsv --input_type fastq --bowtie2out /data/home/btx143/shotgun2/demo/outputs/demo_output/demo_humann_temp/demo_metaphlan_bowtie2.txt
and
Error message returned from metaphlan :
No MetaPhlAn BowTie2 database found (–index option)!
Expecting location /data/home/btx143/.conda/envs/biobakery3/lib/python3.7/site-packages/metaphlan/metaphlan_databases/mpa_v30_CHOCOPhlAn_201901
Exiting…
The issue is that when I check in the directory where the database indices should be, then I find the following files:
mpa_latest
mpa_v30_CHOCOPhlAn_201901.1.bt2
mpa_v30_CHOCOPhlAn_201901.2.bt2
mpa_v30_CHOCOPhlAn_201901.3.bt2
mpa_v30_CHOCOPhlAn_201901.4.bt2
mpa_v30_CHOCOPhlAn_201901.fna
mpa_v30_CHOCOPhlAn_201901.fna.bz2
mpa_v30_CHOCOPhlAn_201901.md5
mpa_v30_CHOCOPhlAn_201901.pkl
mpa_v30_CHOCOPhlAn_201901.rev.1.bt2
mpa_v30_CHOCOPhlAn_201901.rev.2.bt2
mpa_v30_CHOCOPhlAn_201901.tar
And the expected “mpa_v30_CHOCOPhlAn_201901” is missing
I found a similar post in this forum and tried the following:
1. manually delete metaphlan/bowtie2 databases and re-run demo
From what I unnderstand, humann (through metaphlan and bowtie2) should automatically
2. manually delete metaphlan/bowtie2 databases, download the files and use bowtie2 to build the databases
tar -xvf mpa_v30_CHOCOPhlAn_201901.tar
bunzip2 *.bz2
bowtie2-build mpa_v30_CHOCOPhlAn_201901.fna mpa_v30_CHOCOPhlAn_201901
And even tried to run humann specifying the databases option:
humann -i demo.fastq -o sample_results --metaphlan-options “–bowtie2db ${condapath}/${dbpa} -x mpa_v30_CHOCOPhlAn_201901”
After this, I still get the same annoying error:
No MetaPhlAn BowTie2 database found (–index option)!
Expecting location /data/home/btx143/.conda/envs/biobakery3/lib/python3.7/site-packages/metaphlan/metaphlan_databases/mpa_v30_CHOCOPhlAn_201901
Exiting…
Did anyone run into the same or similar error?
Cheers,
Giacomo