MAGMa is a abbreviation for 'Ms Annotation based on in silico Generated Metabolites'. MAGMa subproject which performs actual calculations.
The MAGMa command-line script can be executed in a Docker container. For more information on installing and running Docker see https://www.docker.com/. With the Docker image it is particularly straightforward to run the "light" subcommand which takes a single MS/MS spectrum as input, retrieves candidate structures from the online databases or a local file and writes the result as smiles or sdf to stdout.
Examples:
$ docker run --rm nlesc/magma light -h
$ cat >glutathione.mgf
BEGIN IONS
TITLE=CASMI 2014, Challenge 9
PEPMASS=308.0912 100.0
116.0165 3.2
144.0114 6.3
162.0219 40.2
179.0485 100.0
233.0590 21.6
290.0802 5.1
END IONS
^d
$ docker run --rm -v $PWD:/data nlesc/magma light -f mgf -s hmdb glutathione.mgf# Install conda
# For python 2:
wget https://repo.continuum.io/miniconda/Miniconda2-latest-Linux-x86_64.sh
sh Miniconda2-latest-Linux-x86_64.sh
# Or, for python 3:
wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh
sh Miniconda3-latest-Linux-x86_64.sh
# Optionally, create a dedicated conda environment and activate
conda create -n magma
source activate magma
# Install dependencies
conda install -c rdkit rdkit
conda install cython lxml nose coverage
# For python 2:
pip install http://www.parallelpython.com/downloads/pp/pp-1.6.4.zip
# For python 3:
pip install https://www.parallelpython.com/downloads/pp/pp-1.6.4.4.zip
# If needed install C compiler
sudo apt-get update && sudo apt-get install gcc
# Install MAGMa
git clone https://github.com/NLeSC/MAGMa.git
cd MAGMa/job
python setup.py developA 'magma_job.ini' config file is read from users home directory (~/).
Exampe config file to read candidate molecules from the emetabolomics server:
[magma job]
# Retrieve candidate molecules from the emetabolomics server
structure_database.online = True
structure_database.service = http://www.emetabolomics.org/magma/moleculesExample config file to read candidate molecules from local databases (can be created by the scripts in MAGMa/pubchem):
[magma job]
# Location of structure database from which to retrieve candidate molecules locally
structure_database.online = False
structure_database.pubchem = /home/user/magma_databases/Pubchem_MAGMa.db
structure_database.pubchem_halo = /home/user/magma_databases/Pubchem_MAGMa_halo.db
structure_database.kegg = /home/user/magma_databases/Pubchem_MAGMa_kegg.db
structure_database.kegg_halo = /home/user/magma_databases/Pubchem_MAGMa_kegg_halo.db
structure_database.hmdb = /home/user/magma_databases/HMDB_MAGMa.db
# MACS authentication, used for sending progress reports to MAGMa web application
macs.id = <MAC key identifier>
macs.key = <MAC key>Annotate a tree file using PubChem database:
echo '353.087494: 69989984 (191.055756: 54674544 (85.029587: 2596121, 93.034615: 1720164, 109.029442: 917026, 111.045067: 1104891 (81.034691: 28070, 83.014069: 7618, 83.050339: 25471, 93.034599: 36300, 96.021790: 8453), 127.039917: 2890439 (57.034718: 16911, 81.034706: 41459, 83.050301: 35131, 85.029533: 236887, 99.045074: 73742, 109.029404: 78094), 171.029587: 905226, 173.045212: 2285841 (71.013992: 27805, 93.034569: 393710, 111.008629: 26219, 111.045029: 339595, 137.024292: 27668, 155.034653: 145773), 191.055725: 17000514), 353.087097: 4146696)' > example.tree
magma read_ms_data --ms_data_format tree -l 5 -a 0 example.tree results.db
magma annotate -p5 -q0 -c0 -d0 -b3 -i -1 -s pubchem -o ../pubchem/Pubchem_MAGMa_new.db,0,9999 -f results.dbOn the compute node not all dependencies of Magma will be installed. By freezing the magma application on the head node we include all dependencies like rdkit.
On head node:
pip install bbfreeze
python setup.py bdist_bbfreeze
cd dist
chmod +x dist/Magma-<version>/Magma-<version>-py2.7.egg/magma/script/reactor
tar -zcf Magma-<version>.tar.gz Magma-<version>On compute node:
tar -zxf Magma-<version>.tar.gz
./Magma-<version>/magma ...