Peptide Property Predictor

A unified interface for peptide physicochemical analysis and functional prediction.

Tool Information
About

Enter one peptide sequence per line, select the analyses you need, then submit for prediction.

Example
AGTKDATLPR MKWVTFISLL LLFSSAYSRG
Example Files
Text (.txt) FASTA (.fasta) CSV (.csv)
Instructions
  • One sequence per line
  • Use standard amino acid one-letter codes
  • Allowed length: 5–100 aa
  • Maximum 50 sequences per run
Prediction Models
Sequence Input

Paste sequences directly or load them from a file.

Max 50 sequences Length 5–100 aa
Physicochemical Properties

Property selection.

Click to expand ▶
Structural Properties

Structure-related selection.

Click to expand ▶
Functional Properties

Biological activity and therapeutic potential predictions.

Click to expand ▶
Antibacterial
11 items ▶
Antiviral
7 items ▶
Toxicity
6 items ▶
Therapeutic
9 items ▶
Other Functions
8 items ▶
Antifungal
3 items ▶
No models selected yet.
ESM-2 + GATv2 Model Architecture

All function prediction models share a unified backbone: ESM-2 protein language model produces per-residue embeddings, which are processed by GATv2 graph attention layers to capture long-range residue interactions. Bundle models (Toxicity / Antiviral) output multiple labels simultaneously; single-label models are optimised per functional category.

ESM-2 + GATv2 model architecture