Abstract

In the field of bio-informatics, the analysis of voluminous data is becoming increasingly crucial to understanding the underlying biology and answering important questions. Various clustering techniques such as Hierarchical, SOMs, K-means and PCA are being used to cluster gene expression data to find the functions of unknown genes. Even these sophisticated algorithms are futile if the results are not appropriately interpreted, thus visualization techniques play an important role in analyzing data. Similar to clustering of gene expression data is that of clustering the characteristics of protein folding trajectory and a good visualization tool can help visual analysis and can provide faster and deeper insights into the manner in which a protein folds. With protein characteristics data and specific visualization requirements provided by Dr. Laxmi Parida and Dr. Ruhong Zhou of the Computation Biology Group at the IBM T. J. Watson Research Center, a new 3D visualization technique was designed and developed. This customized technique helps identify the major states a protein folds into through the use of an animated terrain. This technique was implemented as part of the interactive visualization program PROTERAN and tested with the β-Hairpin clustered data provided.