classify seqs V1.19

mothur bugs
1

Hi,

I’m having some trouble running classify.seqs in V 1.19. I’ve downloaded the latest template files etc. and keep getting the same error:

Reading in the ~/Documents/dbs/Silva.bacteria/silva.bacteria.silva.tax taxonomy… DONE.
Generating search database… DONE.
It took 10 seconds generate search database.
Calculating template taxonomy tree… DONE.
Calculating template probabilities… DONE.
It took 56 seconds get probabilities.
Classifying sequences from ~/work/bissettbac.gold.fas …
1965 0
Mismatch with taxonomy and template files. Cannot continue.

At this error MOTHUR quits to the system command prompt. I tried to run classify with various combinations of silva reference and different taxonomies and with the RDP training set and taxonomy, but always got the same error.

Any help / ideas /explanations would be much appreciated,

Oh, I’m running mothur on linux 64-bit, lot’s of RAM.

Thanks,

Andrew

2

What are you supplying as reference=??? it should be silva.bacteria.fasta…

3

HI Pat,

I’ve now used template= / reference=
silva.bacteria.fasta

nogaps.bacteria.fasta - as described in the manual

and the RDP trainingset files (ref=RDPetc.fasta and tax=rdpetc.tax)

I’ve also tried it using “template” (as indicated in the online pages) and “reference” as indicated in your mail, but still no luck.

4

The nogaps should be fine. Can you post the exact commands you are typing in and that are failing?

5

mothur > classify.seqs(fasta=~/work/bissettbac.gold.fas, reference=~/Documents/dbs/Silva.bacteria/nogap.bacteria.fasta, taxonomy=~/Documents/dbs/Silva.bacteria/silva.bacteria.silva.tax)

Reading in the ~/Documents/dbs/Silva.bacteria/silva.bacteria.silva.tax taxonomy… DONE.
Generating search database… DONE.
It took 11 seconds generate search database.
Calculating template taxonomy tree… DONE.
Calculating template probabilities… DONE.
It took 57 seconds get probabilities.
Classifying sequences from ~/work/bissettbac.gold.fas …
Mismatch with taxonomy and template files. Cannot continue.


I've used also your suggested silva.bacteria.fasta as reference, and the rdp references and taxonomies and used both the term "reference" and "template" (which is actually correct? It doesn't seem to make any difference). The result is always the same, "Mismatch with taxonomy and template files. Cannot continue" and mothur quits back to the system command prompt.

Thanks Andrew

6

Hmm, that’s weird. I wonder whether the files got corrupted after you downloaded them. You might try again or get the RDP training set (http://www.mothur.org/w/images/4/49/RDPTrainingSet.zip) that is compatible with mothur. BTW, “reference” is the correct option name.

7

I’ve been playing around with this on the weekend and the problem seems to be in the use of the ~ as a shortcut for /home/user/ (commonly used with linux). When I run the classify seqs command with the full path it runs fine.


I don't really understand why this is, but I now remember I had a similar problem with a previous version and using multiple processors (if I used the ~ it would complete the task on however many processors I specified, but the results files wouldn't be concatenated into one file, I'd end up with as many files as I used processors. This wasn't happening with all commands, just some (but I can't remember which). I guess the upshot is mothur doesn't deal well with the ~ and we should use the full path every time............

is that correct, or am I missing something else here?

Cheers,

Andrew

8

sorry, just re-read my post. The “~” should be a tilda (as in: “~/” = tilda/ = home) but for some reason, although it looks fine, when I push submit it’s changed to a -

Hope it now makes some sense

9

I met the same problem with classify.seqs. Yes, the ‘~’ can cause bug with this command. Use the absolute address instead.