DeterminePPI: Inferring protein-protein interactions by parsimony principle

• Version 1.0
• Last updated: February 16, 2006

Reference#

• Xiang-Sun Zhang, Rui-Sheng Wang, Ling-Yun Wu, Guang-Xu Jin, Luonan Chen. Inferrring protein-protein interactions by parsimony principle. Manuscript.

Method#

DeterminePPI aims to infer protein-protein interaction by parsimony principle. Specifically, inferring domain interactions is formulated as an integer linear programming and solved by relaxed linear programming. In contrast to probabilistic algorithms, DeterminePPI is an algorithm based on a deterministic model.　

Procedure#

DeterminePPI-CV.exe: filename number1 number2

This software is to test our model by cross-validation, where filename is the input file containing training set, domain information. Its format can be referred in the examples in DeterminePPI-CV.zip. number1 is cross-validation number and number2 is the number of non-interacting protein pairs we add into the training set. For our model, we set number2 as 0. That is, we only exploy the given training set and do not add artificial data.

DeterminePPI-noCV.exe: filename1 filename2 number1 number2 filename3

Where filename1 is the given protein interactions and filename2 is the corresponding domain information. number1 is the number of protein interactions and number2 is the number of domains. filename3 is the predicted protein interactions.　

Software#

This is a beta version of the program for preliminary testing. The program is still under development.

• DeterminePPI-CV.rar
• DeterminePPI-noCV.rar