To search, Click below search items.


All Published Papers Search Service


Parallel Geometric Hashing Algorithm for Protein Tertiary Structure


Khalid Jaber, Rosni Abdullah, Fazilah Othman


Vol. 7  No. 7  pp. 126-129


The amount of protein structure has become very huge nowadays, which has led to the need for improving algorithms to cope with this exponential increase. At the same time, processors have become more powerful and affordable with low price. By making use of these advantages, we can construct a powerful parallel system that will overcome the problem of exponential increase of data. This paper presents a Parallel Geometric Hashing Algorithm which will parallelize the sequential Geometric Hashing Algorithm on a cluster. This will lead to powerful, accurate and fast results especially in search and matching. This parallelized algorithm performs coordinate transformations on the feature points of an object to obtain an abstract model of that object. The matching objects are identified by computing correlations between the feature of the query and the model. Our system was run on the Stealth cluster and it used the multiple instruction multiple data (MIMD) paradigm which was applied on protein tertiary structure data. The system allows rapid recognition of unknown protein structure which consists of many residues. We achieved improvement for parallel time speedup compared to the sequential time speedup.


Geometric Hashing Algorithm, Parallel Geometric Hashing Algorithm, model-based recognition, Protein Tertiary Structure