Build the application

## Build
Parse multi-transcript FASTA and metadata packages from NCBI Datasets to detect exon truncation anomalies. Generate structured FHIR Molecular Observation resources linked to ResearchStudy identifiers for curator sign-off.

## User
Bioinformatics Data Curator

## Workflow
1. Ingest zip-compressed gene packages containing RNA transcripts and RefSeq metadata reports.
2. Execute Biopython sequence validation to calculate sequence lengths and identify terminal codon bounds.
3. Map validated transcript properties into FHIR Observation and Molecular DiagnosticReport schema structures.
4. Surface discrepancies in a reconciliation dashboard for human biocuration approval.

## Interface
Web dashboard with transcript length distribution plots, sequence mismatch highlight tables, and FHIR export preview panes.

## Data
Download open NCBI Datasets gene package for human BRCA1 (Gene ID 672) using CLI sandbox mock fixtures.
Starter examples below are synthetic. They are not downloaded records.

## Source
[NCBI Datasets gene download](https://www.ncbi.nlm.nih.gov/datasets/docs/v2/how-tos/genes/download-gene-data-package/)
Uses the NCBI Datasets gene data package archive layout containing transcript RNA fasta and gene metadata json reports.

## Tools
anthropic-claude, biopython, fastapi.
anthropic-claude structures non-deterministic annotation discrepancies into compliant FHIR Molecular Observation JSON payloads.
Hosted Claude API for contextual biological summary synthesis and schema mapping; Biopython running locally for sequence processing.

## Deliverables
Submit runnable source, pinned dependencies, setup commands, a working interface, example outputs, and a recorded end-to-end demo. Include tests and an action log.

## Acceptance
- Parse transcript fasta headers and calculate sequence length for 100% of input records.
- Construct valid FHIR DiagnosticReport JSON containing observation member references.
- Flag transcript sequences with non-canonical termination codons into a review queue.

## Starter Examples
```json
[
  {
    "input": {
      "gene_id": "672",
      "symbol": "BRCA1",
      "transcripts": [
        {
          "accession": "NM_007294.4",
          "length_bp": 7088,
          "terminal_codon": "TAA"
        }
      ]
    },
    "expected_output": {
      "resourceType": "Observation",
      "status": "final",
      "code": {
        "text": "Transcript Length Verification"
      },
      "valueQuantity": {
        "value": 7088,
        "unit": "bp"
      },
      "curation_status": "approved"
    }
  },
  {
    "input": {
      "gene_id": "672",
      "symbol": "BRCA1",
      "transcripts": [
        {
          "accession": "NM_007294.4_trunc",
          "length_bp": 120,
          "terminal_codon": "CCA"
        }
      ]
    },
    "expected_output": {
      "resourceType": "Observation",
      "status": "preliminary",
      "code": {
        "text": "Transcript Length Verification"
      },
      "valueQuantity": {
        "value": 120,
        "unit": "bp"
      },
      "curation_status": "flagged_non_canonical_termination"
    }
  }
]
```

## Safety
Research-only benchmarking pipeline using public synthetic identifiers; excludes patient clinical health information and diagnostic claims.
Use mock external writes. Require human review before any real deployment.