Genie3

What it does

Genie3 is an algorithm for inferring a network from multifactorial data (though other data types seem to work well too). In general Genie3 seems to do a good job with whatever data you can give it. Genie3 uses decision trees to infer the network.

Data

• Any kind of data, but knockout/knockdown/multifactorial data is preferred.

Parameters

k

The number of genes to use for splits. Default is "sqrt" which is based on the sqrt of the number of genes. Other values are "all" and any number from 1 to num_genes

num_trees

The number of trees to generate (default: 500).

tree_method

Method: Extra Trees (value: "ET"), or Random Forests (value: "RT"). Extra Trees is the default and seems to work better in most situations.

Example call

genie3job = GENIE3()
genie3job.setup(data, settings, "MF")
jobman.queueJob(genie3job)