---
title: Validation & QA
description: Preflight Biotite inputs — FASTA headers, non-empty sequences, valid residue codes — then record counts, skipped files and RMSD stats in a SUMMARY.json.
---

Validation lets you catch bad inputs before a run and record a QA summary afterward — preflighting sequences and structures, checking headers and residue codes, and capturing counts and RMSD statistics. Deciding what to skip versus fail fast on is what most affects whether a batch degrades gracefully or corrupts its outputs.

> **Where this code runs**

> You cannot open a shell on the platform yourself. The scripts on this page are plain Biotite Python and shell: ask Ivy in chat to run them in the chat's [code workspace](/docs/sandbox), which ships Biotite, or run them on your own machine.

## Preflight checks

```bash
python - <<'PY'
from biotite.sequence.io import load_sequence
from biotite.structure.io import load_structure
from pathlib import Path

def validate_seq(path: Path) -> bool:
    try:
        load_sequence(str(path), format=None)
        return True
    except Exception:
        return False

def validate_struct(path: Path) -> bool:
    try:
        load_structure(str(path))
        return True
    except Exception:
        return False

seq_ok = [p for p in Path("inputs/sequences").glob("*.fasta") if validate_seq(p)]
struct_ok = [p for p in Path("inputs/structures").glob("*.pdb") if validate_struct(p)]

print("Sequences OK:", len(seq_ok))
print("Structures OK:", len(struct_ok))
PY
```

## QA summary JSON

```python
import json
from pathlib import Path

summary = {
    "biotite_version": "auto",
    "inputs": {
        "sequences": len(list(Path("inputs/sequences").glob("*.fasta"))),
        "structures": len(list(Path("inputs/structures").glob("*.pdb"))),
    },
    "outputs_dir": "outputs/run-latest",
}

Path("outputs").mkdir(exist_ok=True)
with open("outputs/SUMMARY.json", "w", encoding="utf-8") as handle:
    json.dump(summary, handle, indent=2)
```

## Checks to include

- Verify FASTA headers and non-empty sequences.
- Ensure residue codes are valid before alignment/superposition.
- Record skipped files with reasons; fail fast on corrupted archives.
- Capture RMSD/length statistics in the summary for spot checks.