bio-immunoinformatics-immunogenicity-scoring

$npx mdskill add GPTomics/bioSkills/bio-immunoinformatics-immunogenicity-scoring

Score neoantigens by combining MHC binding, processing, and expression factors.

  • Prioritize vaccine candidates using multi-factor immunogenicity models.
  • Depends on MHCflurry for binding and processing predictions.
  • Calculates weighted composite scores from domain-informed factors.
  • Outputs ranked epitope lists for vaccine design selection.

SKILL.md

.github/skills/bio-immunoinformatics-immunogenicity-scoringView on GitHub ↗
---
name: bio-immunoinformatics-immunogenicity-scoring
description: Score and prioritize neoantigens and epitopes for immunogenicity using multi-factor models combining MHC binding, processing, expression, and sequence features. Rank candidates for vaccine design. Use when prioritizing epitopes for vaccine development or identifying the most immunogenic neoantigens.
tool_type: python
primary_tool: mhcflurry
---

## Version Compatibility

Reference examples tested with: MHCflurry 2.1+, numpy 1.26+, pandas 2.2+

Before using code patterns, verify installed versions match. If versions differ:
- Python: `pip show <package>` then `help(module.function)` to check signatures

If code throws ImportError, AttributeError, or TypeError, introspect the installed
package and adapt the example to match the actual API rather than retrying.

# Immunogenicity Scoring

**"Rank my neoantigen candidates by immunogenicity"** → Score and prioritize epitopes using multi-factor models combining MHC binding, proteasomal processing, expression level, and sequence foreignness for vaccine candidate selection.
- Python: `mhcflurry` for binding + processing predictions, custom scoring pipeline

## Multi-Factor Scoring

**Goal:** Calculate a composite immunogenicity score from multiple weighted factors (binding, agretopicity, processing, expression, clonality, foreignness).

**Approach:** Score each factor on a 0-1 scale, then combine via weighted sum with domain-informed weights.

```python
import pandas as pd
import numpy as np

def calculate_immunogenicity_score(peptide_data):
    '''Calculate composite immunogenicity score

    Factors considered:
    1. MHC binding affinity (IC50)
    2. Agretopicity (MT vs WT binding ratio)
    3. Proteasomal processing
    4. TAP transport
    5. Expression level
    6. Clonality (VAF for neoantigens)
    7. Self-similarity (avoid tolerance)

    Each factor scored 0-1, then weighted and combined.
    '''
    scores = {}

    # 1. Binding affinity (lower IC50 = better)
    # Transform to 0-1: 1 at 0nM, 0 at 5000nM
    ic50 = peptide_data.get('ic50_nM', 500)
    scores['binding'] = 1 - min(ic50 / 5000, 1)

    # 2. Agretopicity (MT binds better than WT)
    # Ratio of WT/MT IC50, capped at 10
    agretopicity = peptide_data.get('agretopicity', 1.0)
    scores['agretopicity'] = min(agretopicity / 10, 1)

    # 3. Processing score (from MHCflurry)
    processing = peptide_data.get('processing_score', 0.5)
    scores['processing'] = processing

    # 4. Expression (log scale, capped)
    expression = peptide_data.get('expression_tpm', 10)
    scores['expression'] = min(np.log10(expression + 1) / 3, 1)

    # 5. Clonality (for neoantigens)
    vaf = peptide_data.get('vaf', 0.5)
    scores['clonality'] = vaf

    # 6. Self-similarity (lower = better, less tolerance)
    self_sim = peptide_data.get('self_similarity', 0.5)
    scores['foreignness'] = 1 - self_sim

    # Weighted combination
    weights = {
        'binding': 0.25,
        'agretopicity': 0.20,
        'processing': 0.10,
        'expression': 0.15,
        'clonality': 0.15,
        'foreignness': 0.15
    }

    total = sum(scores[k] * weights[k] for k in weights)

    return total, scores
```

## Processing Prediction

**Goal:** Predict proteasomal cleavage and TAP transport probability for candidate peptides.

**Approach:** Use MHCflurry's Class1ProcessingPredictor to score peptide processing likelihood.

```python
from mhcflurry import Class1ProcessingPredictor

def predict_processing_score(peptides):
    '''Predict proteasomal cleavage and TAP transport

    Processing score reflects probability that peptide will be:
    1. Cleaved from protein by proteasome
    2. Transported by TAP into ER
    3. Loaded onto MHC

    Higher processing score = more likely to be presented
    '''
    predictor = Class1ProcessingPredictor.load()

    results = []
    for peptide in peptides:
        # Need surrounding sequence context for processing
        # In practice, extract from protein context
        pred = predictor.predict(peptides=[peptide])
        results.append({
            'peptide': peptide,
            'processing_score': pred['processing_score'].values[0]
        })

    return pd.DataFrame(results)
```

## Self-Similarity Assessment

**Goal:** Determine whether a candidate peptide resembles self-peptides, indicating potential T-cell tolerance.

**Approach:** Compute pairwise sequence identity against a proteome peptide set and flag high-similarity matches.

```python
def calculate_self_similarity(peptide, proteome_peptides, threshold=0.8):
    '''Check if peptide is similar to self-peptides

    High similarity to self-peptides suggests:
    - T-cells may be tolerized (deleted during development)
    - Lower likelihood of immune response

    Threshold 0.8 = 80% identity considered "self-like"
    '''
    def sequence_identity(seq1, seq2):
        if len(seq1) != len(seq2):
            return 0
        matches = sum(1 for a, b in zip(seq1, seq2) if a == b)
        return matches / len(seq1)

    max_similarity = 0
    most_similar = None

    for self_peptide in proteome_peptides:
        sim = sequence_identity(peptide, self_peptide)
        if sim > max_similarity:
            max_similarity = sim
            most_similar = self_peptide

    return {
        'similarity': max_similarity,
        'is_self_like': max_similarity >= threshold,
        'closest_self': most_similar
    }
```

## Hydrophobicity at Position 2

**Goal:** Assess MHC anchor residue quality by checking hydrophobicity at key positions.

**Approach:** Check whether position 2 and C-terminal residues fall within the hydrophobic amino acid set preferred by HLA-A*02:01-like alleles.

```python
def check_anchor_hydrophobicity(peptide):
    '''Check hydrophobicity at MHC anchor positions

    For HLA-A*02:01 and similar alleles:
    - Position 2: Prefers hydrophobic (L, I, V, M)
    - Position 9 (C-terminus): Prefers hydrophobic (L, V, I)

    Strong anchors improve binding stability.
    '''
    hydrophobic = set('LIVMFYW')

    pos2 = peptide[1] if len(peptide) > 1 else ''
    pos_last = peptide[-1]

    return {
        'pos2_hydrophobic': pos2 in hydrophobic,
        'pos_last_hydrophobic': pos_last in hydrophobic,
        'anchor_score': (pos2 in hydrophobic) + (pos_last in hydrophobic)
    }
```

## Rank Epitopes

**Goal:** Rank epitopes by composite immunogenicity score and assign confidence tiers for prioritization.

**Approach:** Score all candidates, sort by immunogenicity, and assign high/medium/low tiers based on percentile ranking.

```python
def rank_epitopes(epitope_df, top_n=20):
    '''Rank epitopes by immunogenicity

    Returns top candidates with scores and confidence tiers.

    Confidence tiers:
    - High: Top 5%, all factors favorable
    - Medium: Top 20%, most factors favorable
    - Low: Remaining, some factors favorable
    '''
    epitope_df = epitope_df.copy()

    # Calculate scores
    scores = []
    factor_scores = []
    for _, row in epitope_df.iterrows():
        total, factors = calculate_immunogenicity_score(row.to_dict())
        scores.append(total)
        factor_scores.append(factors)

    epitope_df['immunogenicity_score'] = scores
    factor_df = pd.DataFrame(factor_scores)

    # Combine
    result = pd.concat([epitope_df, factor_df], axis=1)

    # Rank
    result = result.sort_values('immunogenicity_score', ascending=False)

    # Assign tiers
    n = len(result)
    result['tier'] = 'low'
    result.iloc[:int(n * 0.20), result.columns.get_loc('tier')] = 'medium'
    result.iloc[:int(n * 0.05), result.columns.get_loc('tier')] = 'high'

    return result.head(top_n)
```

## Compare Candidates

**Goal:** Select a diverse set of vaccine candidates with broad HLA coverage and non-overlapping positions.

**Approach:** Iterate through ranked candidates, selecting those with non-overlapping genomic positions to maximize epitope diversity.

```python
def compare_vaccine_candidates(candidates_df):
    '''Compare and select vaccine candidates

    Vaccine design typically selects:
    - Multiple epitopes (5-20)
    - Diverse HLA coverage
    - High immunogenicity scores
    - Non-overlapping sequences
    '''
    # Group by HLA coverage
    hla_coverage = candidates_df.groupby('allele').size()

    # Select diverse set
    selected = []
    used_positions = set()

    for _, candidate in candidates_df.iterrows():
        # Check for overlap with selected
        pos = candidate.get('position', 0)
        if not any(abs(pos - p) < 5 for p in used_positions):
            selected.append(candidate)
            used_positions.add(pos)

        if len(selected) >= 20:
            break

    return pd.DataFrame(selected)
```

## Related Skills

- immunoinformatics/mhc-binding-prediction - Binding affinity component
- immunoinformatics/neoantigen-prediction - Input candidates
- immunoinformatics/epitope-prediction - Epitope identification

More from GPTomics/bioSkills

SkillDescription
bio-admet-predictionPredicts ADMET properties using ADMETlab 3.0 API or DeepChem models. Estimates bioavailability, CYP inhibition, hERG liability, and 119 toxicity endpoints with uncertainty quantification. Filters for PAINS and other structural alerts. Use when filtering compounds for drug-likeness or prioritizing leads by predicted safety.
bio-alignment-amplicon-clippingTrim PCR primers from aligned reads in amplicon-panel BAMs using samtools ampliconclip. Use when processing SARS-CoV-2 ARTIC, hereditary cancer panels, ctDNA hot-spot panels, or any amplicon assay where primer-derived bases would falsely confirm reference at primer footprints.
bio-alignment-filteringFilter alignments by flags, mapping quality, and regions using samtools view and pysam. Use when extracting specific reads, removing low-quality alignments, or subsetting to target regions.
bio-alignment-indexingCreate and use BAI/CSI indices for BAM/CRAM files using samtools and pysam. Use when enabling random access to alignment files or fetching specific genomic regions.
bio-alignment-ioRead, write, and convert multiple sequence alignment files using Biopython Bio.AlignIO. Supports Clustal, PHYLIP, Stockholm, FASTA, Nexus, and other alignment formats for phylogenetics and conservation analysis. Use when reading, writing, or converting alignment file formats.
bio-alignment-msa-parsingParse and analyze multiple sequence alignments using Biopython. Extract sequences, identify conserved regions, analyze gaps, work with annotations, and manipulate alignment data for downstream analysis. Use when parsing or manipulating multiple sequence alignments.
bio-alignment-msa-statisticsCalculate alignment statistics including sequence identity, conservation scores, substitution matrices, and similarity metrics. Use when comparing alignment quality, measuring sequence divergence, and analyzing evolutionary patterns.
bio-alignment-multiplePerform multiple sequence alignment using MAFFT, MUSCLE5, ClustalOmega, or T-Coffee. Guides tool and algorithm selection based on dataset size, sequence divergence, and downstream application. Use when aligning three or more homologous sequences for phylogenetics, conservation analysis, or evolutionary studies.
bio-alignment-pairwisePerform pairwise sequence alignment using Biopython Bio.Align.PairwiseAligner. Use when comparing two sequences, finding optimal alignments, scoring similarity, and identifying local or global matches between DNA, RNA, or protein sequences.
bio-alignment-sortingSort alignment files by coordinate or read name using samtools and pysam. Use when preparing BAM files for indexing, variant calling, or paired-end analysis.