EMAIL:  PASSWORD:      

About GAT

Gene-set Activity Toolbox or GAT

This is a web-based application, which provides several computational tools and materials for supporting the gene expression study including, the simple analysis tool to identify genetic markers, the gene-set activity transformation to analyze the data more systematically, the machine learning tools from WEKA as well as we also provide the 15 benchmark gene-set activity datasets.

Fig.1 shows the workflow of GAT. Initially, this tool allows to analyze gene expression data from Gene Expression Omnibus (GEO) only. After retriving the dataset from GEO, the process will be started from pre-process, gene-set activity transformation, the classification to the biological intepretation of results.


Citing us

- Engchuan, W., Chan, J. H.: Pathway activity transformation for multi-class classification of lung cancer datasets. Neurocomputing, in press (2014), doi:10.1016/j.neucom.2014.08.096

- Sootanan, P., Prom-on, S., Meechai, A., & Chan, J. H.: Pathway-based microarray analysis for robust disease classification. Neural Computing and Applications, 21(4), 649-660 (2012)


Fig.1 Overview of GAT.

Simple Gene Expression Analysis

GAT provide the simple tools to visualize the gene expression data, identify differential expression genes for the preliminary study step.
***Not available yet!

Gene-set Activity Transformation

Instead of analyzing gene expression levels, many work have been done in converting them to another form namely Pathway activity or Gene-set activity by integrating with pathway or gene-set data. The pathway/gene-set activity have been successfull applied for disease classification. Here, we provide several gene-set activity transformation methods including CORR-based, NCFS-i, AFS, etc.

Machine Learning

This tool is integrated with WEKA (the data mining tool) to maximize the learning effectiveness from the data. This includes the feature selection, classification, clustering and model validation with various number of methods.

Intepretation of Result

After the machine learning processes is done, the results from those step would be more valuable, if it can be annotated its relationship to the phenotype outcome. In this version, we only allow user to link and intepret their result using KEGG pathway database via KEGG mapper.

***Currently undergoing construction.

Data Repository

The benchmark gene-set activity datasets are available to download for who interesting in using these benchmark datasets to evaluate their novel algorithms for analyzing these gene-set activities. The gene-set activity datasets available >> here <<.
Home   |   Data Repository   |   Java Library   |   Contact Us and Developers   |  
© 2015 Gene-set Activity Toolbox, School of Information Technology, King Mongkut's University of Technology Thonburi Contact person: Assoc.Prof.Dr. Jonathan H. Chan (jonathan (at) sit.kmutt.ac.th)