Learn R Programming

predictionet (version 1.18.0)

Inference for predictive networks designed for (but not limited to) genomic data

Description

This package contains a set of functions related to network inference combining genomic data and prior information extracted from biomedical literature and structured biological databases. The main function is able to generate networks using Bayesian or regression-based inference methods; while the former is limited to < 100 of variables, the latter may infer networks with hundreds of variables. Several statistics at the edge and node levels have been implemented (edge stability, predictive ability of each node, ...) in order to help the user to focus on high quality subnetworks. Ultimately, this package is used in the 'Predictive Networks' web application developed by the Dana-Farber Cancer Institute in collaboration with Entagen.

Copy Link

Version

Version

1.18.0

License

Artistic-2.0

Homepage

http://compbio.dfci.harvard.edu

Maintainer

Benjamin HaibeKains

Last Published

February 15th, 2017

Functions in predictionet (1.18.0)

jorissen.colon.ras

Gene expression, annotations, clinical data and priors for the colon cancer tumors collected by Jorissen and colleagues in 2009.

Function to make prediction of a node values given its parents using an inferred network

predictionet.press.statistic

Function computing the press statistic for all target variables in topology

Function to identify all children of a parent

Function performing network inference by combining priors and genomic data

adj.remove.cycles

Function to remove cycles that may be present in a directed graph represented by an adjacency matrix

predictionet.stability.cv

Function inferring networks in cross-validation

Function computing performance of prediction; methods include r2, nrmse and mcc

Function computing the f1-score, comparing an inferred topology with a given topology

predictionet-package

Inference for predictive networks designed for (but not limited to) genomic data

Function to create an igraph object and export a network to a GML readable by Cytoscape

Gene expression, annotations, clinical data and priors for the colon cancer tumors collected by the expression project for oncology (expO).

Function performing network inference by combining priors and genomic data

Function fitting a regression model for each gene in the data

data.discretize

Function to discretize data based on user specified cutoffs

Function to compute the Matthews Correlation Coefficient (MCC) in a classification framework