Neural Network Model for Prediction of PPI using Domain Frequency & Association Score Base Classification of Protein Pairs

Main Article Content

Garima Srivastava
Navita Srivastava, Gulshan Wadhwa


Developing In Silico Computational Techniques to predict Protein Protein Interactions (PPIs) is one of the challenging research area for computational biologists. Experimental techniques for interaction prediction, lack accuracy and are error prone. Observing the limitations and low accuracy of existing methods, this work focuses on improvement in computational efficiency with increased performance for domain base prediction of PPIs. In the present paper, a novel approach of Domain Frequency Count (DFC) method with association score base classification, using feed forward back-propagation neural network, has been proposed for prediction of interacting protein pairs based on their domain characteristic features data. Results obtained are quite encouraging. When compared with the existing MLE, DT, CL NN, and RDF techniques [8][10] on similar datasets, it is observed that accuracy and sensitivity are increased by 8.91% and 5.70%, respectively.

Keywords: Neural Network, Protein-Protein Interaction, Domain-Domain Interaction, Domain Frequency Count, Association Score.


Download data is not yet available.

Article Details