description.md

Data Sources and Methods

Training Data Collection

Our models are trained on curated datasets from multiple sources:

Hemolysis Dataset

• Primary Source: the Database of Antimicrobial Activity and Structure of Peptides (DBAASPv3)
• Secondary Source: peptideBERT
• Size: 9,316 peptides, with 19.6% being positive (hemolytic) and 80.4% being negative (nonhemolytic)
• Description: Probability of peptide disrupting red blood cell membranes.
• Download: hemo-positive.npz
• Interpretation 50% of read blood cells being lysed at x ug/ml concetration (HC50). If HC50 < 100uM, considered as hemolytic, otherwise non-hemolytic.

Solubility Dataset

• Primary Source: PROSO-II
• Secondary Source: peptideBERT
• Size: 18,453 sequences, with 47.6% being labeled as positives and 52.4% being labeled as negatives
• Description: Probability of peptide remaining dissolved in aqueous conditions.
• Download: sol-positive.npz

Non-Fouling Dataset

• Primary Source: Classifying antimicrobial and multifunctional peptides with Bayesian network models
• Secondary Source: peptideBERT
• Size: 3,600 positive, 13,585 negative
• Description: A nonfouling peptide resists nonspecific interactions and protein adsorption.
• Download: nf-positive.npz

Permeability Dataset

• Primary Source: CycPeptMPDB, PAMPA
• Secondary Source: PepLand
• Size: 1162 positive and negative for nanonical samples each (22 relevant cell-penetrating peptide databases by compiling literature on existing cell-penetrating peptide prediction models ); CycPeptMPDB provides extra 7334 cyclic peptides
• Description: Probability of peptide penetrating the cell membrane.
• Download: nc-CPP-processed.csv

Half-life Dataset

• Primary Source: Thpdb2, PepTherDia, peplife
• Size: 105 wt, 275 wt+noncanonical, human-only
• Clean-ups: Data are all transformed into log(hour)
• Download: binding_affinity_training_data.csv

Binding Affinity Dataset

• Primary Source: PepLand
• Size: 1,781 protein-peptide complexes, canonical and non-canonical
• Description: Binding probability normalized in PepLand already. It's a combination of IC50/EC50.
• Quality: Binding class cutoffs: Tight ≥ 7.5, Medium 6.0–7.5, Weak < 6.0
• Download: binding_affinity_training_data.csv
• Interpretation Affinity_measure = -log_10(K).

Model Architecture

• Sequence Embeddings: ESM-2 650M parameter model
• XGBoost Models: Gradient boosting on pooled ESM embeddings
• CNN Models: 1D convolutional networks with attention mechanisms
• Binding Model: Cross-attention between protein and peptide representations

Model Training and Weight Hosting

• Classifier_weights

🧪 Physicochemical Properties

Net Charge Calculation

• Uses Henderson-Hasselbalch equation
• pH-dependent calculation
• Considers all ionizable groups (K, R, H, D, E, C, Y, termini)

Isoelectric Point (pI)

• Bisection method to find pH where net charge = 0
• Precision: ±0.01 pH units

Hydrophobicity (GRAVY)

• Grand Average of Hydropathy
• Uses Kyte-Doolittle scale
• Range: -4.5 (hydrophilic) to +4.5 (hydrophobic)

Citation

If you use this tool, please cite:

place holder

Contact

For questions or collaborations: yzhang@u.duke.nus.edu