AMETHST, or Analysis METHod Selection Tool, is an automated pipeline that makes it possible to objectively assess the relative performance of multiple analysis methods applied to annotation abundance data. Utilizing objective, permutation-based statistics, it can determine the best performing analysis from any arbitrary collection of analyses, and can also determine the degree to which each analysis step (e.g. normalization and reprocessing of the data, selection of distance/dissimilarity metric) affects the results produced by each analysis method.
configure deploy.cfg
make
one of:
make deploy-client
make deploy-service
make deploy-backend
make deploy-all
then
source /kb/deployment/user-env.sh
one of
make test-client
make test-service
make test-backend
make test
Script location (the script is located in the AMETHST repository linked to this amethst_service repository):
AMETHST/installation/Install_AMETHST_compute_node.havannah.sh
Script should be executed with envrionment variables that include KB_AUTH_TOKEN and AWE_CLIENT_GROUP_TOKEN:
sudo -E Install_AMETHST_compute_node.havannah.sh
USAGEPOD library
wget https://raw.githubusercontent.com/wgerlach/USAGEPOD/master/lib/USAGEPOD.pm
SHOCK client library
mkdir SHOCK
cd SHOCK
wget https://raw.githubusercontent.com/MG-RAST/Shock/master/libs/SHOCK/Client.pm
cd ..
amethst client script
wget https://raw.githubusercontent.com/kbase/amethst_service/master/plbin/mg-amethst.pl
chmod +x mg-amethst.pl
configuration to use KBase amethst service
export SHOCK_SERVER_URL=
export KB_AUTH_TOKEN=<your kbase token>
The token can also be passed with --token= to mg-amethst.pl