Triplet-SVM and MiPred are ab initio tools for predicting whether an RNA sequence is a precursor micro RNA (pre-miRNA). They rely on local structural features of the RNA sequences, and use machine learning methods to handle the classification. This provides a snazzy web interface to both of them, using node.js and socket.io to provide a less clunky experience than is typical.
You will need the following dependencies:
-
R, and the randomForest package for it.
-
node.js, version 0.3.1 or higher.
-
The socket.io and paperboy packages for node.js, which you can install with npm.
Once you have all these, make sure that the svm-classify
program
from libsvm is in your PATH
, and run the server:
$ node server.js
Then just point your web browser at http://localhost:8080/ and enjoy!
This is no longer being actively maintained. If you have a use for it, and you would like to take over the project, just fork it on GitHub. You have my blessing.
Peter Scott,
December 2010