MHCII3D

MHCII3D predicts peptides that bind Major Histocompatibility Complex class II (MHC II) molecules using structure-based modeling and optimized statistical scoring to support immunological research.

Key Features:

• Structure-Based Approach: Utilizes structural scaffolds derived from known MHC II-peptide complexes to model three-dimensional peptide–MHC II interactions.
• Statistical Scoring Functions (SSFs): Employs optimized SSFs to evaluate binding affinity and stability of antigen sequences threaded into scaffold models.
• 3D Modeling Pipeline: Processes MHC II alleles from the Immuno Polymorphism Database (IPD) to generate distinct scaffold complexes for each allotype sequence.
• Sequence Threading: Threads antigen protein sequences through structural scaffolds of MHC II-peptide complexes to produce modeled interactions.
• Database Curation Detection: Identifies problematic entries in databases such as the Immune Epitope Database (IEDB).
• Robustness Evaluation: Demonstrates resilience against overfitting via leave-one-out cross-validation tests.

Scientific Applications:

• Complementary Prediction Methodology: Integrates structural modeling with SSFs to complement sequence-based machine learning methods trained on experimentally characterized binding peptides.
• Benchmarking: Benchmarked against sequence-based methods using Immune Epitope Database (IEDB) datasets, achieving a Pearson correlation coefficient of 0.42 with experimentally determined IC50 values.
• Immunological Research Support: Supports identification and validation of potential peptide binders relevant to cancer, autoimmune diseases, and allergies.

Methodology:

Antigen protein sequences are threaded through structural scaffolds of known MHC II-peptide complexes generated by a 3D-modeling pipeline from IPD alleles and evaluated using optimized statistical scoring functions (SSFs), with identification of problematic IEDB entries.