Loading...

Portfolio

Some of my work

Attention Based Convolutional Neural Network For Predicting Protein–Protein Interactions

  • Date:
    Late Winter 2019
  • Coding Language:
    Jupyter Notebook, Python, Matlab
  • Dependencies:
    Keras 1.1.2, Theano 0.9.0 Sklearn, Numpy, Scikit-learn
  • Operating system:
    Linux (Ubuntu 17.04)
  • Organisation:
    Indian Institute of Technology Patna (IIT Patna)
  • Supervisor:
    Prof. Dr. Sriparna Saha
Description

A physical contact is established in protein-protein interaction (PPI) from more than one protein molecules. In the biocuration process, PPI extraction is a process for relation extraction . We have build an attention based convolutional neural network to discover protein-protein interaction.We have tested on a pair of labeled protein entities which are passing through a sentence G. If it founds the interaction between two entities then the output will be true. To encode the contextual information, a convolution layer is used. In our method we used attention mechanisms to get all information about the target relation by intuitively reconciling the relational proportionality of every phrase and interaction and learning the sentence representation.

  • Project Image