LearnBoo is a MATLAB software with a graphical user interface (GUI) for learning the Boolean transition functions given a signed directed network and experimental observations about the state of nodes. The software provides network visualization of the learned networks, and allows users to simulate in-silico gene/protein/node knockdowns under the learned Boolean framework.
LearnBoo can be freely downloaded from this web-site. LearnBoo requires MATLAB to be installed. It was developed and tested under a Windows OS with the MATLAB 2010 version. In addition, the MATLAB Bioinformatics Toolbox is required to be installed on the host computer.
Running the LearnBoo software: To run LearnBoo GUI double-click the learnBoo.fig file or type in learnBoo in the MATLAB command window. The screen below should appear:
Uploading data onto LearnBoo:Input data includes: (1) an adjacency matrix with element Aij representing a link from variable/node/gene/protein i to j; (2) a data matrix with rows representing conditions and columns representing variables/nodes/genes/proteins and (3) a file with variables/nodes/genes/proteins name annotations. Example dataset can be downloaded from here. The example input data should be unzipped and saved in the same directory as the LearnBoo MATLAB files.
Data Analysis (learning a model from the data): After the input data was imported, users should first perform the data binarization step in the Analysis panel if the data is not already in binary format. The following image shows the results from the binarization process:
The next step is to learn the Boolean logic of the directed network. This is achieved by simply clicking the Learn Boolean Logic button. The learned networks could be visualized in another MATLAB figure window by clicking the Draw Network button once the learning process completed. Learned transition functions can be exported to user-defined destinations by clicking the Export button.
Data Analysis (running simulations with perturbations): In addition, users can perform dynamical simulations of single and combinatorial perturbations of the network nodes and obtain the representative final state of the network after simulations under the Perturbation panel.
Please, contact us if you have any questions or suggestions.