OMA error - download from gene ontology not working 404

Question

[nc52]

Hi there,

Wondering if anyone can help with an OMA error I am having. It seems that OMA cannot access the file it requires from www.geneontology.org/ontology/gene_ontology.obo This is the error message:

Starting database conversion and checks...
Process 38697 on marvin: job nr 1 of 100
--2019-02-20 10:58:07--  http://www.geneontology.org/ontology/gene_ontology.obo
Resolving www.geneontology.org... 107.23.42.138
Connecting to www.geneontology.org|107.23.42.138|:80... connected.
HTTP request sent, awaiting response... 301 Moved Permanently
Location: http://geneontology.org/ontology/gene_ontology.obo [following]
--2019-02-20 10:58:07--  http://geneontology.org/ontology/gene_ontology.obo
Resolving geneontology.org... 107.23.42.138
Reusing existing connection to www.geneontology.org:80.
HTTP request sent, awaiting response... 302 Found
Location: http://purl.obolibrary.org/obo/go/gene_ontology.obo [following]
--2019-02-20 10:58:07--  http://purl.obolibrary.org/obo/go/gene_ontology.obo
Resolving purl.obolibrary.org... 52.3.123.63
Connecting to purl.obolibrary.org|52.3.123.63|:80... connected.
HTTP request sent, awaiting response... 302 Found
Location: http://snapshot.geneontology.org/ontology/gene_ontology.obo [following]
--2019-02-20 10:58:07--  http://snapshot.geneontology.org/ontology/gene_ontology.obo
Resolving snapshot.geneontology.org... 143.204.178.196, 143.204.178.95, 143.204.178.64, ...
Connecting to snapshot.geneontology.org|143.204.178.196|:80... connected.
HTTP request sent, awaiting response... 404 Not Found
2019-02-20 10:58:07 ERROR 404: Not Found.

Error, (in wget) wget failed

I have confirmed with our cluster manager that queue I am submitting to has internet access and that this problem is not due to a firewall at our end.

I have also noticed that if you copy www.geneontology.org/ontology/gene_ontology.obo directly into a browser you do indeed get a 404 not found. To manually download the file you need to start at www.geneontology.org and navigate through the website to the .obo.

I tried placing the (manually downloaded) .obo in the directory where the script was launched and the /DB and parameters.drw are located in case OMA would look for it there before attempting to download but that didnt work.

Has anyone experienced this problem or have any thoughts about a fix? Any help greatly appreciated.

Answer 1

[Christophe Dessimoz]

Thanks for reporting this problem. It indeed appears that the Gene Ontology consortium website was recently updated and the URL has changed. We will fix this and do a new release soon. Meanwhile, if you (or anyone else) is affected, here is a fix:

Locate the "darwinlib" directory. If you are running OMA standalone directly from the decompressed archive, it is in that directory. If you have installed it locally, it might be in the directory "/usr/local/OMA/OMA.2.3.0/". If you installed it using Homebrew, it will be in a place like "/usr/local/Cellar/oma/2.3.0/OMA/OMA.2.3.0//".

In that directory, edit the file called "Ontology" and update the URL from line 60 to

http://purl.obolibrary.org/obo/go.obo

[nc52]

Thank you very much for the prompt reply and fix! Seems to be working :D Much appreciated!

[nc52]

Dear Christophe,

It seems as though I spoke too soon! We are now getting the following error:

Starting database conversion and checks... Process 64271 on marvin: job nr 1 of 100 Reading GO file... Ontologies read: 47360 Error, (in ConvertRawFile) Darwin cannot open file /root/.oma/GOdata.drw-201902\ 25 to write to: Permission denied

Could you please advise as to what part of the code to alter to put output file to an area which the user will have permissions to do so?

Many thanks, Nicki

[Adrian Altenhoff]

We fixed this problem in OMA 2.3.1. Please upgrade to at least this version if you have this problem

Answer 2

[Christophe Dessimoz]

OMA standalone relies on the shell utility "mktemp", which should create a temporary file that is writable for the process, but somehow this seems to be failing in your system. This might have to do with how your job scheduler is configured on your cluster.

Could you try to run a simple script which calls mktemp and tries to write to it?

#!/bin/sh
TMPFILE=`mktemp`
echo "$TMPFILE"
echo "test" >> $TMPFILE

If this does not work, please contact your cluster system administrator to report the issue, and please let us know what sort of scheduler caused this problem.

Meanwhile, as a workaround, you can replace line 58 of the Ontology file, which starts with "rawfile := ", by

rawfile := "nomktemp";

which is what gets triggered in systems which have no mktemp installed.