Protein-protein Interaction Prediction using Arabic semantic analysis

摘要

Scientists are still far from unraveling the molecular mechanisms of most relevant diseases such as cancer and diabetes. A better understanding of protein interactions could provide a clue about the molecular mechanism of the processes leading to such diseases. Novel methodologies to understand diseases through their primary protein interactions are highly desired. In this paper we propose a simple method to predict protein-protein interaction based on Arabic semantic analysis model. The Arabic semantic model is an effective feature extraction method based on natural language processing. Two protein sequences may interact if they contain similar or related Arabic words. The semantic meaning will most likely provide us with a clue on how or why two proteins interact. To evaluate the ability of the proposed method to distinguish between “interacted” and “non-interacted” proteins pairs, we applied it on a dataset of 200 protein pairs from the available yeast saccharomyces cerevisiae protein interaction. The proposed method managed to get 100% sensitivity, 0.84% sensitivity and 92% overall accuracy. The method also showed moderate improvement over the existing well-known methods for PPI prediction such as PPI-PS and PIPE.