Databases not indexed; after indexing they *still* aren't indexed? Or are they missing?

[wolfgangrumpf]

When I run Prokka with:

Prokka scaffolds.fasta

I get this:

[15:34:32] This is prokka 1.14.5
[15:34:32] Written by Torsten Seemann <[email protected]>
[15:34:32] Homepage is https://github.com/tseemann/prokka
[15:34:32] Local time is Thu Dec  3 15:34:32 2020
[15:34:32] You are ubuntu
[15:34:32] Operating system is linux
[15:34:32] You have BioPerl 1.7.7
Argument "1.7.7" isn't numeric in numeric lt (<) at /usr/bin/prokka line 259.
[15:34:32] System has 4 cores.
[15:34:32] Option --cpu asked for 8 cores, but system only has 4
[15:34:32] Will use maximum of 4 cores.
[15:34:32] Annotating as >>> Bacteria <<<
[15:34:32] The sequence databases have not been indexed. Please run 'prokka --setupdb' first.

I ran Prokka --setupdb and the Prokka --listdb and got:

[15:35:08] Looking for databases in: /var/lib/prokka/db
[15:35:08] * Kingdoms:
[15:35:08] * Genera: Enterococcus Escherichia Staphylococcus
[15:35:08] * HMMs:
[15:35:08] * CMs: Archaea Bacteria Viruses

Running Prokka scaffolds.fasta just gives me the same error (databases not indexed).

What am I doing wrong?

[wolfgangrumpf]

I commented out the code that does the database check (lines 324-328) and ran Prokka again; it seems that the databases really aren't set up - I got this on my next run attempt:

14:00:06] Running: cat PROKKA_12042020\/PROKKA_12042020\.sprot\.tmp\.146140\.faa | parallel --gnu --plain -j 8 --block 110403 --recstart '>' --pipe blastp -query - -db /var/lib/prokka/db/kingdom/Bacteria/sprot -evalue 1e-09 -qcov_hsp_perc 80 -num_threads 1 -num_descriptions 1 -num_alignments 1 -seg no > PROKKA_12042020\/PROKKA_12042020\.sprot\.tmp\.146140\.blast 2> /dev/null
[14:00:06] Could not run command: cat PROKKA_12042020\/PROKKA_12042020\.sprot\.tmp\.146140\.faa | parallel --gnu --plain -j 8 --block 110403 --recstart '>' --pipe blastp -query - -db /var/lib/prokka/db/kingdom/Bacteria/sprot -evalue 1e-09 -qcov_hsp_perc 80 -num_threads 1 -num_descriptions 1 -num_alignments 1 -seg no > PROKKA_12042020\/PROKKA_12042020\.sprot\.tmp\.146140\.blast 2> /dev/null

[kapsakcj]

How did you install prokka? From your second output, it appears that none of the 'kingdoms' or 'HMM' databases were set up, despite 'Genera' and 'CMs' databases being set up. A bit odd.

Here's what I see in the docker container for prokka that I use:

$ prokka --listdb
[15:05:09] Looking for databases in: /prokka-1.14.5/db
[15:05:09] * Kingdoms: Archaea Bacteria Mitochondria Viruses
[15:05:09] * Genera: Enterococcus Escherichia Staphylococcus
[15:05:09] * HMMs: HAMAP
[15:05:09] * CMs: Archaea Bacteria Viruses

Perhaps try prokka --cleandb and then prokka --setupdb --dbdir /var/lib/prokka/db to give it the specific location of the databases. It might have just been a blip when setting up the databases.

[mmp3]

bump! I have the same exact problem as @wolfgangrumpf I installed prokka on ubuntu 20.04.1 LTS via apt-get install prokka

I tried the commands given by @kapsakcj , i.e.

prokka --cleandb
prokka --setupdb --dbdir /var/lib/prokka/db
prokka  --dbdir /var/lib/prokka/db  genome.fa

but the error remained.

[kapsakcj]

I think @mr-c is the one of the maintainers for the prokka apt package (apologies for tagging you if not)? He may be able to assist @mmp3 and @wolfgangrumpf if that's also how they installed. This changelog indicates that the current prokka apt package for Ubuntu 20 is unstable.

You may get set up quicker if you pursue a different route of install via (bio)conda, brew, docker, install from source, etc.

[mmp3]

I just installed via the method outlined in the README file for ubuntu:

sudo apt-get install libdatetime-perl libxml-simple-perl libdigest-md5-perl git default-jre bioperl
sudo cpan Bio::Perl
git clone https://github.com/tseemann/prokka.git $HOME/prokka
$HOME/prokka/bin/prokka --setupdb

but the problem remains.

Note: I indeed uninstalled the version from the apt package manager and called the version of prokka cloned from the git repository explicity:

$HOME/prokka/bin/prokka --setupdb
$HOME/prokka/bin/prokka  genome.fa

and I got the same error.

[mr-c]

Hello everyone, please use reportbug when you have an issued with a Debian/Ubuntu package as it gathers the requisite information and routes it to the right volunteers.

This changelog indicates that the current prokka apt package for Ubuntu 20 is unstable.

FYI: 'unstable' is the name of the Debian development distribution. If there is no change needed, then this changelog entry is copied as the package moves to Debian's testing and stable distributions; likewise with Ubuntu and other Debian derivaties

[kapsakcj]

Thanks for chiming in, Michael. And thank you for the info and context on what 'unstable' means! Much appreciated.

[mmp3]

Thank you, @mr-c . Please note thouh, @kapsakcj , that the error persists even with the version of prokka from github, as demonstrated above.

[kapsakcj]

@mmp3 I recommend using conda, brew, or docker/singularity instead of installing from source + apt. Could you try one of those routes?

[Pixelka91]

@wolfgangrumpf Hey I ran into the same issue. Tried all of the above, but it didn't work. What worked thou, was copying the DB files from $HOME/prokka/db folder to /var/lib/prokka/db
Still don't know why it stopped working in the first place. I annotated genomes with PROKKA a couple of months before without any issues.

[admontecillo0428]

This works!

@wolfgangrumpf Hey I ran into the same issue. Tried all of the above, but it didn't work. What worked thou, was copying the DB files from $HOME/prokka/db folder to /var/lib/prokka/db Still don't know why it stopped working in the first place. I annotated genomes with PROKKA a couple of months before without any issues.

[F-anwer]

please change your directory where you placed your genome fasta files