Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jan 10, 2026

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

ParaDeep: sequence-based deep learning for residue-level paratope prediction using chain-aware BiLSTM-CNN models.

Piyachat Udomwong¹, Thanathat Pamonsupornwichit², Kanchanok Kodchakorn^3,4

¹International College of Digital Innovation, Chiang Mai University, Chiang Mai, Thailand.

Frontiers in Bioinformatics

|November 21, 2025

Summary

Related Concept Videos

Per-Unit Sequence Models

Per-Unit Sequence Models

An ideal Y-Y transformer, grounded through neutral impedances, displays per-unit sequence networks akin to those of a single-phase ideal transformer when subjected to balanced positive- or negative-sequence currents. These currents do not produce neutral currents, and their associated voltage drops.
Zero-sequence currents, which are identical in magnitude and phase, generate a neutral current, resulting in voltage drops across the neutral impedance and the low-voltage winding. If the...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Experimental and Mechanistic Validation of PARP1pred for Identifying Potent Leads.

Computational and structural biotechnology journal·2026

Same author

Fc engineering of a fully humanized anti-CD147 monoclonal antibody enhances ADCC against T-cell acute lymphoblastic leukemia and T-lymphoblastic lymphoma.

Scientific reports·2026

Same author

Interferon-gamma autoantibodies in patients with infectious uveitis: a pilot case-control study.

BMC infectious diseases·2026

Same author

Development of a monoclonal antibody-based competitive ELISA as a surrogate assay for detecting neutralizing anti-interferon gamma autoantibodies in adult-onset immunodeficiency.

PloS one·2026

Same author

Integrating deep learning with physics based modeling enables high precision antibody antigen interface prediction.

Scientific reports·2026

Same author

Comparative evaluation of E. coli expression systems for soluble hScFv-IL-17A with an unpaired CDR-L3 cysteine.

Journal, genetic engineering & biotechnology·2025

ParaDeep accurately predicts antibody paratopes from amino acid sequences using deep learning. This lightweight framework offers high performance, especially for heavy chains, accelerating antibody discovery without structural data.

Area of Science:

Computational biology
Immunoinformatics
Machine learning in drug discovery

Background:

Accurate antibody paratope prediction is crucial for antibody discovery workflows.
Existing methods often rely on structural data, limiting high-throughput applications.

Purpose of the Study:

To develop ParaDeep, a deep learning framework for residue-level paratope prediction directly from antibody amino acid sequences.
To evaluate the performance of ParaDeep across various model configurations and antibody chain types.

Main Methods:

ParaDeep integrates bidirectional long short-term memory (LSTM) networks with 1D convolutional layers.
Systematic evaluation of 30 model configurations, varying encoding schemes, kernel sizes, and antibody chain types (heavy and light).
Performance assessed using five-fold cross-validation and an independent blind test set, employing F1 score and Matthews Correlation Coefficient (MCC).

Keywords:

BiLSTM-CNN antibody binding site prediction deep learning heavy and light chains paratope identification sequence modeling

Related Experiment Videos

Last Updated: Jan 10, 2026

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

Main Results:

Heavy chain models achieved superior performance (F1 = 0.856, MCC = 0.842) in cross-validation compared to light chain models (F1 = 0.774, MCC = 0.772).
On the blind test set, ParaDeep achieved F1 = 0.723 and MCC = 0.685 for heavy chains, outperforming the baseline Parapred by 27% in MCC.
Heavy chains provide stronger sequence-based predictive signals, while light chains benefit more from structural context.

Conclusions:

ParaDeep offers a lightweight, interpretable, and sequence-based approach for paratope prediction, rivaling structure-based methods for heavy chains.
Its efficiency and scalability are ideal for early-stage antibody discovery, repertoire profiling, and therapeutic design, especially when structural data is unavailable.
The framework's accessibility via open-source code and Google Colab facilitates broader adoption in antibody research.