Immunogenicity Class 1¶
Predicts MHC class I immunogenicity for input protein structures by scanning sequences with a sliding peptide window and scoring predicted binding to a specified class I allele. Returns a CSV table of immunogenicity scores for all peptides/windows per structure.
Usage¶
Use this node to pre-screen designed proteins for potential MHC class I immunogenic peptides. Provide one or more PDB structures, select the target allele (e.g., HLA-A0101), and set the peptide window size (typically 8–11, default 9). Integrate it after structure generation/selection to filter candidates before downstream analysis or experimental validation.
Inputs¶
| Field | Required | Type | Description | Example |
|---|---|---|---|---|
| pdb | True | PDB | One or more protein structures to evaluate. Each entry should be a named structure representing a single protein for which peptide immunogenicity will be assessed. | {'designA_ranked_0.pdb': 'PDB_STRING_CONTENT', 'designB_ranked_0.pdb': 'PDB_STRING_CONTENT'} |
| c_allele | True | STRING | MHC class I allele against which immunogenicity (peptide binding) is predicted. | HLA-A0101 |
| c_window_size | True | INT | Sliding window size (peptide length) used to generate peptides for binding prediction. Class I peptides are typically 8–11 residues; default is 9. | 9 |
| mode | True | ['MOCK', 'PROD', 'TEST'] | Execution mode. MOCK returns predefined example results, PROD runs the actual prediction service, TEST uses minimal parameters for quick validation (not biologically meaningful). | PROD |
Outputs¶
| Field | Type | Description | Example |
|---|---|---|---|
| score.csv | CSV | CSV table of immunogenicity scores for peptides derived from the input structures. Typically includes multiple rows per structure (one per peptide window) with scores and identifiers. | protein_id,peptide,start,end,allele,score example_protein,SIINFEKL,10,17,HLA-A0101,0.87 |
Important Notes¶
- Allele format: For class I, use allele strings like HLA-A0101 (no colons).
- Peptide length: Class I predictions are most relevant for 8–11-mers; default is 9.
- Multiple rows per structure: Output contains one row per peptide window, so expect many rows per input protein.
- Performance: Processing time scales with the number of structures and the window size.
- TEST mode behavior: In TEST mode, window size is automatically reduced to 3 and only the first structure is processed. Results are for validation only.
- MOCK mode: Returns canned data useful for UI and pipeline testing; not for scientific interpretation.
Troubleshooting¶
- Empty or missing scores: Verify that the PDB inputs contain valid, parseable polypeptide chains and that c_window_size is appropriate (e.g., 9).
- Unexpectedly short peptides: Ensure mode is set to PROD (not TEST), as TEST reduces peptide length for speed.
- Unsupported or invalid allele: If the run fails or returns errors, check that c_allele matches supported naming (e.g., HLA-A0101).
- Timeouts with many inputs: Reduce the number of structures or use a smaller window size, or run in batches to avoid timeouts.
- CSV appears duplicated: The output lists multiple peptide windows per protein. This is expected when scanning across the sequence.
Example Pipelines¶
