HUMAnN3 (No MetaPhlAn BowTie2 database found (–index option)!)

Question

[a.mostafa5050]

Hi,

I am trying to run Humann3, I downloaded the databases:

humann_databases --download chocophlan full /cluster/projects/nn8021k/humann_database
humann_databases --download uniref uniref90_diamond /cluster/projects/nn8021k/humann_database

Upon running the human, I got that error (No MetaPhlAn BowTie2 database found (–index option)!)

I googled the issue but no solution.

Could you please help me with that?

[cpad0112]

Did you try giving location manually (--nucleotide-database for chocophlan and --protein-database for uniref)? Look at the output from 'humann_config --print' command. This would give you the locations available for humann3. You can also update the path by running humann_config --update and supplying appropriate values for the databases. If the databases are in place, make sure that they are latest databases. Please read the manual.

[a.mostafa5050]

Thanks for the reply :)

I have just downloaded the database yesterday, however, I updated them again today.

humann_databases --download chocophlan full /cluster/projects/nn8021k/humann_database/chocophlan --update-config yes
humann_databases --download chocophlan full /cluster/projects/nn8021k/humann_database/uniref --update-config yes
humann_databases --download utility_mapping full /cluster/projects/nn8021k/Conda-env/humann/lib/python3.9/site-packages/humann/data/misc --update-config yes

Then I ran my job, but I got the same error.

Error message returned from metaphlan :
No MetaPhlAn BowTie2 database found (--index option)!
Expecting location bowtie2db
Exiting...

This is my command line

humann --input /cluster/projects/nn8021k/11-unmapped_fastq_zip/zipp-files/dataset-for-metaphlan/sorted_bamfilt_decont_trimmed_1-1_cat.fastq.gz --nucleotide-database /cluster/projects/nn8021k/humann_database/chocophlan --protein-database /cluster/projects/nn8021k/humann_database/uniref --output /cluster/projects/nn8021k/Humann_work/OUTPUT

[Ram]

Please do not add answers unless you're answering the top level question. Use Add Comment or Add Reply as appropriate.

I've moved your answer to a reply to the appropriate comment.

[cpad0112]

can you try running following command and check if humann3 is picking up correct databases?

humann_config --print'

Try running metaphlan independently first on your data. If this fails, then you cannot run humann3 as humann runs metaphlan first internally.

[a.mostafa5050]

Yes, the path to the databases is correct after running (humann_config --print)

But still got the same error.

When I ran Metaphlan independently, it used a different database (mpa_v30_CHOCOPhlAn_201901).

After running the human using this database, it is said it is an old version.

Should I index the databases independently using bowtie2 or bwa?