Detection of unannotated protein functions in a protein interaction network generates a lot of beneficial information in the field of drug discovery of various kinds of diseases. Though most of the various computational methods have succeeded in predicting functions of huge amount of unknown proteins at recent times but the main problem is the simultaneous increase of false positives in most of the predicted results. In this work, a bottom-up predictor of existing Apriori algorithm has been implemented for protein function prediction by exploiting two most important neighborhood properties: closeness centrality and edge clustering coefficient of protein interaction network. The method is also unique in the fact that the functions of the leaf nodes in the interaction network have been back propagated and thus labeled up to the root node (target protein) using a bottom-up level to level approach. An overall precision, recall and F-score of 0.86, 0.65 and 0.74 respectively have been obtained in this work which are found to be better than most of the current state-of-the-art.
CITATION STYLE
Prasad, A., Saha, S., Chatterjee, P., Basu, S., & Nasipuri, M. (2017). Protein Function prediction from protein interaction network using bottom-up L2L apriori algorithm. In Communications in Computer and Information Science (Vol. 776, pp. 3–16). Springer Verlag. https://doi.org/10.1007/978-981-10-6430-2_1
Mendeley helps you to discover research relevant for your work.