humann2_rename_table problem with uniref90

[valeri...@uniroma1.it]

I have tried to run the script rename_table as follows:
humann2_rename_table -i S1_merged_PE_genefamilies.tsv -o prova -n uniref90

this yielded the following error:

humann2_rename_table: error: argument -n/--names: invalid choice: 'uniref90' (choose from 'uniref50', 'kegg-module', 'kegg-orthology', 'kegg-pathway', 'metacyc-rxn', 'ec', 'metacyc-pwy')

I am really surprised since uniref90, based on the tutorial I followed, should be supported by humann2_rename_table. What am I missing?

Thanks in advance for your kind help,

Valerio

[Eric Franzosa]

Hi Valerio,

The rename file for UniRef90 was added as of HUMAnN2 v0.8.0. Are you using v0.8.0+?

Thanks,

Eric

[Valerio Fulci]

Hi Eric,

I am using humann2 v 0.9.1

Hard to say which version of Humann2_rename_table I am using since the humann2_rename_table does not seem to support --version option or anything like that.
I have, to the best of my knowledge, installed humann2 only once on my PC, however I double checked, and both humann2.py and humann2_rename_table.py files were created at the same day/time, so I really believe that the only humann2_rename_table command in my path links to a script installed together with the humann2 (version 0.9.1) I am using.

However, to be honest python on my system is slightly messed up since I have installed packages using bot pip and miniconda, and those two guys seem to set environmental variables which are not really compatible. Indeed I ended up with python2.7, python3.4 and python 3.5 installed with parallel implementations of scipy and numpy...

Nevertheless, everything else seems to work nicely, since each python easily and nicely finds its own modules without any issue thus far.

I do not believe that this should generate this kind of error, however, to cross-check I performed today a fresh humann2 install on a PC with a cleaner python install by using pip install humann2 and run the same command on the same table, obtaining the very same error message.

Valerio

Valerio Fulci, PhD
Dipartimento di Biotecnologie Cellulari ed Ematologia
Sapienza Università di Roma
Policlinico Umberto I, V Clinica Medica
Vle Regina Elena, 324
00161 Rome
Italy

Phone: +39064457731
Fax:+39064457731

[Eric Franzosa]

Hi Valerio,

Have you updated the utility mapping recently? 

$ humann2_databases --download utility_mapping full <destination directory>

It's possible you are missing the map from UniRef90 IDs to names, which is required for uniref90 to show up in the list of options for humann2_rename_table.

And indeed, the HUMAnN2 utility tools are tied to a version of HUMAnN2, so your HUMAnN2 version is also the version of the scripts. Having the scripts also report the version directly is a great idea - I will look into it.

Thanks,

Eric

Thanks,

Eric

[florentinc...@gmail.com]

Hello,

I had the same issue.
I fallowed the tutorial using one sample of my dataset to get comfortable with your awesome pipeline.

florentin@S620100019620:~$ humann2 --version
humann2 v0.9.3

I updated the utility mapping, as you suggesteed and it solved the issue.

Cheers,

Flo

[maron...@gmail.com]

Hi Eric.

I ran into the same trouble as Valerio. When I renamed the gene family file with uniref90, it said there is none uniref90 command as follows:

humann2_rename_table: error: argument -n/--names: invalid choice: 'uniref90' (choose from 'infogo1000', 'metacyc-rxn', 'kegg-module', 'ec', 'go', 'metacyc-pwy', 'pfam', 'eggnog', 'uniref50', 'kegg-pathway', 'kegg-orthology')

So I wanted to fix this problem as you recommended, use this command to download the full mapping dataset:

$ humann2_databases --download utility_mapping full <destination directory>

and I succeed in downloading it and installing it, eventually the program has the choice of uniref90. So I was very excited to run my sample but unfortunately it can't get the rename annotation:

Treating Sample_TCM-10.nonmouse_genefamilies.tsv as stratified output, e.g. ['UniRef50_unknown', 'g__Bacteroides.s__Bacteroides_uniformis']
Loading mapping file from: /public/users/chentingting/lib/python2.7/site-packages/humann2/data/misc/utility_mapping/map_uniref90_name.txt.bz2
Renamed 0 of 64962 entries (0.00%)

I was wondering if there is something wrong with my sample so I tested the demo in humann2, but it failed again. I am confused with this kind of problem, could u help me check this kind of problem?

Loading table from: demo_genefamilies.tsv
Treating demo_genefamilies.tsv as stratified output, e.g. ['UniRef50_R5IH84', 'g__Bacteroides.s__Bacteroides_stercoris']
Loading mapping file from: /public/users/chentingting/lib/python2.7/site-packages/humann2/data/misc/utility_mapping/map_uniref90_name.txt.bz2
Renamed 0 of 1413 entries (0.00%)

and the output files shows as below:

# Gene Family demo_Abundance-RPKs
UNMAPPED 100.0000000000
UniRef50_R5IH84: NO_NAME 37.0370370370
UniRef50_R5IH84: NO_NAME|g__Bacteroides.s__Bacteroides_stercoris 37.0370370370
UniRef50_B0NQY6: NO_NAME 25.6410256410
UniRef50_B0NQY6: NO_NAME|g__Bacteroides.s__Bacteroides_stercoris 25.6410256410
UniRef50_F3PHD0: NO_NAME 25.6410256410
UniRef50_F3PHD0: NO_NAME|g__Bacteroides.s__Bacteroides_stercoris 25.6410256410
UniRef50_B0NP96: NO_NAME 23.8095238095
UniRef50_B0NP96: NO_NAME|g__Bacteroides.s__Bacteroides_stercoris 23.8095238095
UniRef50_B0NTS9: NO_NAME 21.2765957447
UniRef50_B0NTS9: NO_NAME|g__Bacteroides.s__Bacteroides_stercoris 21.2765957447
UniRef50_R6FNM5: NO_NAME 16.1290322581
UniRef50_R6FNM5: NO_NAME|g__Bacteroides.s__Bacteroides_thetaiotaomicron 16.1290322581
UniRef50_A6L108: NO_NAME 15.6250000000
UniRef50_A6L108: NO_NAME|g__Bacteroides.s__Bacteroides_stercoris 15.6250000000

Sincerely,
Rong



[Eric Franzosa]

Hi Rong,

Looks like you're working with a UniRef50 table (e.g. UniRef50_R5IH84), so you'll want to use the UniRef50 rename option and not UniRef90.

Thanks,

Eric

[Rong Ma]

Hi Eric,

Thank you for replying me. I found the latest ec filtered database you provided has the annotation for gene family. So I could get annotation without renaming them but should rename those files gotten with old database. Thank u again.

Yours,

Rong

Rong Ma

State Key Laboratory(Ophthalmology), Zhongshan Ophthalmic Center of Sun Yat-sen University 

Address: 403 of 3rd Building,Zhongshan Ophthalmic Center of Sun Yat-sen University ,Yuexiu District,Guangzhou

Tel: +86 155-2114-8964