Deep Learning Approaches for Protein Secondary Structure Prediction

Umme Israt Afroz; Shreshthajit Das; Nazmul Siddique

doi:10.1109/iccit64611.2024.11021832

Deep Learning Approaches for Protein Secondary Structure Prediction

Umme Israt Afroz
, Shreshthajit Das
, Nazmul Siddique

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Citations (Scopus)

Abstract

This paper presents a novel hybrid model comprising Evolutionary Scale Modeling (ESM), Convolution Neural Network (CNN) and Long Short Term Memory (LSTM) network for prediction of protein secondary structures from coil (C), helix (H), and sheet (E)—from amino acid sequences using deep learning techniques. Each architecture leverages unique strengths, with LSTMs capturing long-range dependencies, CNNs extracting local spatial patterns, and ESM enhancing contextual understanding of sequences. The hybrid model was trained and tested using two key datasets: the UniProt dataset and the pdb-intersect-pisces dataset, which provide a rich source of protein sequences and structural information. The proposed model achieved an accuracy of 89.22%, demonstrating robust performance in protein secondary structure prediction.

Original language	English
Title of host publication	2024 27th International Conference on Computer and Information Technology (ICCIT)
Publisher	IEEE
Pages	3474-3479
Number of pages	6
ISBN (Electronic)	9798331519094
DOIs	https://doi.org/10.1109/iccit64611.2024.11021832
Publication status	Published online - 10 Jun 2025
Event	2024 27th International Conference on Computer and Information Technology (ICCIT) - Cox's Bazar, Bangladesh Duration: 20 Dec 2024 → 22 Dec 2024

Publication series

Name	2024 27th International Conference on Computer and Information Technology (ICCIT)
Publisher	IEEE Control Society
ISSN (Print)	2474-9648
ISSN (Electronic)	2474-9656

Conference

Conference	2024 27th International Conference on Computer and Information Technology (ICCIT)
Country/Territory	Bangladesh
City	Cox's Bazar
Period	20/12/24 → 22/12/24

Bibliographical note

Publisher Copyright:
© 2024 IEEE.

Keywords

Proteins
Deep learning
Accuracy
computational modeling
Neural networks
Computer architecture
predictive models
Transformers
information technology
Long short term memory
CNN
LSTM
Prediction
Protein structure
ESM

Access to Document

10.1109/iccit64611.2024.11021832

Cite this

Afroz, U. I., Das, S., & Siddique, N. (2025). Deep Learning Approaches for Protein Secondary Structure Prediction. In 2024 27th International Conference on Computer and Information Technology (ICCIT) (pp. 3474-3479). (2024 27th International Conference on Computer and Information Technology (ICCIT)). IEEE. Advance online publication. https://doi.org/10.1109/iccit64611.2024.11021832

@inproceedings{9f99f7f041ce45a9be459ec71b21113e,

title = "Deep Learning Approaches for Protein Secondary Structure Prediction",

abstract = "This paper presents a novel hybrid model comprising Evolutionary Scale Modeling (ESM), Convolution Neural Network (CNN) and Long Short Term Memory (LSTM) network for prediction of protein secondary structures from coil (C), helix (H), and sheet (E)—from amino acid sequences using deep learning techniques. Each architecture leverages unique strengths, with LSTMs capturing long-range dependencies, CNNs extracting local spatial patterns, and ESM enhancing contextual understanding of sequences. The hybrid model was trained and tested using two key datasets: the UniProt dataset and the pdb-intersect-pisces dataset, which provide a rich source of protein sequences and structural information. The proposed model achieved an accuracy of 89.22\%, demonstrating robust performance in protein secondary structure prediction.",

keywords = "Proteins, Deep learning, Accuracy, computational modeling, Neural networks, Computer architecture, predictive models, Transformers, information technology, Long short term memory, CNN, LSTM, Prediction, Protein structure, ESM",

author = "Afroz, \{Umme Israt\} and Shreshthajit Das and Nazmul Siddique",

note = "Publisher Copyright: {\textcopyright} 2024 IEEE.; 2024 27th International Conference on Computer and Information Technology (ICCIT) ; Conference date: 20-12-2024 Through 22-12-2024",

year = "2025",

month = jun,

day = "10",

doi = "10.1109/iccit64611.2024.11021832",

language = "English",

series = "2024 27th International Conference on Computer and Information Technology (ICCIT)",

publisher = "IEEE",

pages = "3474--3479",

booktitle = "2024 27th International Conference on Computer and Information Technology (ICCIT)",

address = "United States",

}

Afroz, UI, Das, S & Siddique, N 2025, Deep Learning Approaches for Protein Secondary Structure Prediction. in 2024 27th International Conference on Computer and Information Technology (ICCIT). 2024 27th International Conference on Computer and Information Technology (ICCIT), IEEE, pp. 3474-3479, 2024 27th International Conference on Computer and Information Technology (ICCIT), Cox's Bazar, Bangladesh, 20/12/24. https://doi.org/10.1109/iccit64611.2024.11021832

Deep Learning Approaches for Protein Secondary Structure Prediction. / Afroz, Umme Israt; Das, Shreshthajit; Siddique, Nazmul.
2024 27th International Conference on Computer and Information Technology (ICCIT). IEEE, 2025. p. 3474-3479 (2024 27th International Conference on Computer and Information Technology (ICCIT)).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Deep Learning Approaches for Protein Secondary Structure Prediction

AU - Afroz, Umme Israt

AU - Das, Shreshthajit

AU - Siddique, Nazmul

PY - 2025/6/10

Y1 - 2025/6/10

N2 - This paper presents a novel hybrid model comprising Evolutionary Scale Modeling (ESM), Convolution Neural Network (CNN) and Long Short Term Memory (LSTM) network for prediction of protein secondary structures from coil (C), helix (H), and sheet (E)—from amino acid sequences using deep learning techniques. Each architecture leverages unique strengths, with LSTMs capturing long-range dependencies, CNNs extracting local spatial patterns, and ESM enhancing contextual understanding of sequences. The hybrid model was trained and tested using two key datasets: the UniProt dataset and the pdb-intersect-pisces dataset, which provide a rich source of protein sequences and structural information. The proposed model achieved an accuracy of 89.22%, demonstrating robust performance in protein secondary structure prediction.

AB - This paper presents a novel hybrid model comprising Evolutionary Scale Modeling (ESM), Convolution Neural Network (CNN) and Long Short Term Memory (LSTM) network for prediction of protein secondary structures from coil (C), helix (H), and sheet (E)—from amino acid sequences using deep learning techniques. Each architecture leverages unique strengths, with LSTMs capturing long-range dependencies, CNNs extracting local spatial patterns, and ESM enhancing contextual understanding of sequences. The hybrid model was trained and tested using two key datasets: the UniProt dataset and the pdb-intersect-pisces dataset, which provide a rich source of protein sequences and structural information. The proposed model achieved an accuracy of 89.22%, demonstrating robust performance in protein secondary structure prediction.

KW - Proteins

KW - Deep learning

KW - Accuracy

KW - computational modeling

KW - Neural networks

KW - Computer architecture

KW - predictive models

KW - Transformers

KW - information technology

KW - Long short term memory

KW - CNN

KW - LSTM

KW - Prediction

KW - Protein structure

KW - ESM

UR - https://www.scopus.com/pages/publications/105009079965

U2 - 10.1109/iccit64611.2024.11021832

DO - 10.1109/iccit64611.2024.11021832

M3 - Conference contribution

T3 - 2024 27th International Conference on Computer and Information Technology (ICCIT)

SP - 3474

EP - 3479

BT - 2024 27th International Conference on Computer and Information Technology (ICCIT)

PB - IEEE

T2 - 2024 27th International Conference on Computer and Information Technology (ICCIT)

Y2 - 20 December 2024 through 22 December 2024

ER -

Deep Learning Approaches for Protein Secondary Structure Prediction

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this