| |
Code and Usage |
| NEWS |
We are moving the source code to GP-DREAM. Go here for the latest versions . We also have a preliminary version that can be run on GP-DREAM!
|
| Source code |
| NEW: We have made the Hierarchical clustering part of MERLIN faster and less memory expensive. Please download the newest version of MERLIN |
| merlin.tgz After downloading type tar -xvzf merlin.tgz This will create
a directory in the current path called merlin. cd in to the directory and type make. The executable is named learnMERLIN. |
| Third party libraries |
| MERLIN needs GNU scientific library to run. We provide libraries compiled x84 64 bit Linux platform. |
| Usage |
| ./learnMERLIN -m example/in/yeast_expr_interpolated -o example/out/ -r4 -p-5 -h 0.6 -l example/in/regulators.txt -v1 -c example/in/clusterassign.txt -k300 |
| |
There are several arguments merlin takes, however, the most important ones are r, p, h, v, c and l described below. We are working on improving the documentation and usability of merlin :
- -m model
- -o outputdir
- -k maxfactorsize
- -v cross_validation_cnt
- -l restricted_regulator_fname
- -p sparsity_prior
- -r modularity_prior
-
- -c initial_cluster_assignments
|
| UTILITIES |
| MERLIN uses the .data and .model format originally introduced with the Bayesian network learning in PNL (Probabilistic Network Library). We are providing a utility program called merlinconverter
than can be used to convert the expression data matrix into .data and .model format. The directory has a minimal README with usage to show how to use the program |
| |
