PepPharmaHub: a cloud-based platform integrating multimodel language architectures with curated data resources for...

Area of Science:

• Computational biology
• Peptidomics
• Drug discovery

Background:

• Therapeutic peptides are promising drug candidates but predicting their function from sequence is challenging.
• Existing computational tools lack the flexibility and scalability for modern drug discovery workflows.
• A cloud-based, no-code platform is needed for customizable modeling and high-throughput screening of therapeutic peptides.

Purpose of the Study:

• To develop a cloud-based, no-code platform for therapeutic peptide functional prediction.
• To enable customizable deep learning models for high-throughput screening.
• To accelerate peptide-based drug discovery through efficient functional analysis.

Main Methods:

• Developed PepPharmaHub, a cloud-based platform integrating sequence-based language modeling, benchmark datasets, and visualization.
• Implemented a high-throughput screening module with 24 models for 20 therapeutic properties.
• Included a customizable model training pipeline for user-defined screening tasks.

Main Results:

• PepPharmaHub matches or surpasses state-of-the-art predictors on 24 public datasets.
• Demonstrated significantly improved efficiency in large-scale peptide screening.
• Achieved higher, more consistent accuracy on recent bioactive peptides compared to existing web servers.

Conclusions:

• PepPharmaHub provides accurate, high-throughput peptide function prediction using customizable deep learning models.
• The platform's no-code interface and superior performance accelerate therapeutic peptide discovery.
• Offers interpretable sequence analysis, enhancing practical utility for researchers.