Towards User-friendly Interfacing of Biomedical Applications with the Grid: A Paradigm with SVM Optimization for Gene Prediction

Grid computing is currently considered as the key technology towards realization of the e-Science vision, aiming to enable high throughput research. Among several application domains, Grid computing is envisioned to have a strong impact in biomedical research, which is typically characterized by computationally intensive and complex applications. However, although there are significant advances in the construction and availability of the underlying Grid infrastructure worldwide, there is still lack of user-friendly means to either access the Grid or enable the execution of existing biomedical software into Grid infrastructures. In this paper, we present our effort towards the construction of a generic and user-friendly mechanism for accessing and executing applications on the Grid, which requires minor technical knowledge from the end-users and may be used as a paradigm for diverse biomedical applications. The applicability of the proposed UI mechanism is illustrated through a parameter optimization problem of legacy SVMs applied for gene prediction, requiring complex computations and a large amount of training data.