For questions, bug reports, or collaboration inquiries regarding the CAN-IMMUNE database, please contact us via email.

• Mutation Data Collection

  Multi-source cancer mutation curation

  

  Mutation data collected from three primary sources:

  • COSMIC v100: 4,952,684 missense substitutions from 1,460 cancer cell lines
  • CCLE: 968,457 cancer-specific missense substitutions from WES data
  • Published Literature: 806 mutations from 86 additional cell lines

  Total: 6,721,816 cancer-specific mutations across 33 cancer types

• Mutation Validation & Cross-referencing

  Quality control and verification

  

  Mutations cross-referenced with established protein databases:

  • RefSeq Database (GRCh38): Gene names and positions
  • UniProt Database: Amino acid sequences and annotations

  Validation ensures accuracy of mutation annotations including gene location, amino acid changes, and transcript IDs

• Mutant Peptide Library Generation

  Creating searchable libraries

  

  Extraction of mutant peptide sequences:

  • Peptide Length: 25 amino acids (12 residues upstream + mutation + 12 residues downstream)
  • Coverage: Sufficient for HLA class I peptides (8-14 amino acids)
  • Output: 1,194,608 unique mutant peptides from 19,768 genes

  Libraries optimized for MS search engines (FragPipe, PEAKS, DIA-NN)

• CAN-IMMUNE Platform

  Web-based interface and API

  

  Comprehensive web platform features:

  • Browse: Explore mutations by cell line, tissue, or cancer type
  • Search: Query and filter mutations with Elasticsearch
  • Statistics: Interactive visualizations of mutation distributions
  • Download: Export libraries in multiple formats
  • MutPep Tool: Generate custom libraries from user data

  ---

  • Home Page Browse Database Search Mutations Statistics
    MutPep Tool

  Access CAN-IMMUNE database functions
• LC-MS/MS Immunopeptidomics Analysis

  Mass spectrometry-based identification

  

  Comprehensive MS workflow:

  • Sample Preparation: MHC peptide purification from cancer cells
  • LC-MS/MS Acquisition: High-resolution mass spectrometry
  • Database Search: MSFragger/PEAKS with custom mutant libraries
  • FDR Control: Stringent 1% false discovery rate

  Compatible with multiple search platforms for neoantigen discovery

• Peptide Rescoring & Validation

  Enhanced confidence scoring

  

  Advanced rescoring algorithms:

  • Percolator: Statistical FDR control
  • MSBooster: Deep learning-based features
  • MS2Rescore: Peptide identification enhancement
  • PeptideProphet: Probability scoring (>0.9 threshold)

  Reduces false positives in expanded search spaces

• HLA Binding Affinity Prediction

  NetMHCpan 4.1 analysis

  

  Peptide-HLA binding classification:

  • Strong Binders: %EL_rank ≤ 0.5
  • Weak Binders: 0.5 < %EL_rank ≤ 2
  • Non-Binders: %EL_rank > 2

  Supports multiple HLA alleles (Class I: HLA-A, -B, -C)

  Case Study Results: 76% of TNBC mutant peptides predicted as strong binders

• Structural Modeling & TCR Interaction

  AI-powered structure prediction

  

  Advanced computational modeling:

  • PANDORA: Fast peptide-HLA complex modeling
  • AlphaFold2: TCR:peptide-MHC structure prediction
  • Analysis: Binding energy and conformational assessment
  • Comparison: Mutant vs. wild-type structural differences

  Identifies structural alterations affecting T-cell recognition

• Neoantigen Candidate Prioritization

  Ranking for experimental validation

  

  Multi-criteria ranking system:

  • Peptide Quality: PeptideProphet score > 0.9
  • HLA Binding: Strong binder classification
  • Structural Stability: Optimal peptide-HLA conformation
  • TCR Recognition: Favorable interaction interfaces
  • Expression: RNA-seq validation (optional)

  Output: Ranked list of high-confidence neoantigen candidates for immunogenicity testing