Skip to content

Latest commit

 

History

History
14 lines (11 loc) · 781 Bytes

README.md

File metadata and controls

14 lines (11 loc) · 781 Bytes
Raw

protein-secondary-structure-prediction

The aim of this repo is to provide implementation of BRNN model for protein secondary structure prediction. Long-term aim is to improve the current known accuracy for the same task. Done by Udayraj Deshmukh and Sudhanshu Ranjan

Steps to run this model

  1. chmod +x dependencies_installation.sh processCB613.sh
  2. run processCB613.sh
  3. run model_6.py

Results

The model reaches an accuracy of 68.4% on test data set (CB513).

Related work : Next-Step Conditioned Deep Convolutional Neural Networks Improve Protein Secondary Structure Prediction (1702.03865.pdf on arxiv) has an accuracy of 70.3% and also contains comparision with previous models.