Open AccessStacked Bidirectional Long Short Term Memory Models to Predict Protein Secondary Structure
Author(s) -
R. Thendral,
A. N. Sigappi
Publication year - 2020
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.c8368.019320
Subject(s) - computer science , protein secondary structure , sequence (biology) , protein structure database , term (time) , protein structure prediction , class (philosophy) , artificial intelligence , data mining , protein data bank (rcsb pdb) , protein structure , machine learning , biology , sequence database , genetics , biochemistry , physics , quantum mechanics , gene
Protein Secondary Structure (PSS) is one of most complex problem in biology PSS is important for determining tertiary structure in the future, for studying protein fiction and drug design. However, Experimental PSS approaches are time consuming and difficult to implement, and its most essential to establish effective computing methods for predicting on protein sequence structure. Accuracy of prediction performance has been recently improved due to the rapid expansion of protein sequences and the design of libraries in deep learning techniques. In this research proposed a deep recurring network unit method called stacked bidirectional long-term memory (Stacked BLSTM) units to predict 3-class protein secondary structure from protein sequence information using a bidirectional LSTM layer. To evaluate the output of Stacked BLSTM, using publicly available datasets from the RCSB server. This study indicates that performance of our method is better than the of that latest stranded public dataset. The accuracy achieved is more than 89%.