Build a Pydantic AI FHIR Observation Curation Pipeline for NCBI Datasets

Parse multi-transcript FASTA packages downloaded via NCBI Datasets. Evaluate coding sequence integrity and use Pydantic AI to map metrics into strongly typed FHIR R5 Observation JSON bundles.

Bioinformatics Data Curator

Ingest gene data package zip archives containing transcript and CDS sequences.

Parse FASTA sequence records using Biopython SeqIO to verify start codons and length multiples.

Map validated sequence features into FHIR R5 Observation components referencing NCBI gene identifiers using Pydantic AI.

Generate a validated JSON bundle with human-readable diagnostic summaries for cohort curation.

CLI tool that accepts gene FASTA archives and emits compliant FHIR R5 JSON bundles.

Download public gene packages using the NCBI Datasets CLI tool for non-clinical model organisms or reference genes. Starter examples below are synthetic. They are not downloaded records.

NCBI Datasets gene download Utilizes the NCBI Datasets gene data package archive structure containing gene, RNA, CDS, and protein FASTA sequences.

pydantic-ai, biopython. Pydantic AI guarantees type-safe validation when mapping sequence metrics into FHIR JSON schema specifications. Local or hosted LLM with structured output support via OpenAI-compatible or Anthropic API.

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.

Correctly parses multi-record CDS FASTA files and flags invalid frame lengths.

Maps gene IDs and transcripts to FHIR R5 Observation and DiagnosticReport JSON models using Pydantic AI.

Emits complete FHIR Bundles conforming to standard JSON serialization.

```json [ { "input": { "gene_id": "672", "fasta_header": "NM_007294.4 BRCA1 CDS [location=1..228]", "sequence": "ATGGATTTATCTGCTCTTCGCGTTGAAGAAGTACAAAATGTCATTAATGCTATGCAGAAAATCTTAGAGTGTCCCATCTGTCTGGAGTTGATCAAGGAACCTGTCTCCACAAAGTGTGACCACATATTTTGCAAATTTTGCATGCTGAAACTTCTCAACCAGAAGAAAGGGCCTTCACAGTGTCCTTTATGTAAGAATGATATAACCAAAAGGAGCCTACAAGAAAGTACGAGATTTAGTCAACTTGTTGAAGAGCTATTGAAAATCATTTGTGCTTTTCAGCTTGACACAGGTTTGGAGTATGCAAACAGCTATAATTTTGCAAAAAAGGAAAATAACTCTCCTGAACATCTAAAAGATGAAGTTTCTATCATCCAAAGTATGGGCTACAGAAACCGTGCCAAAAGACTTCTTACAGAAAGTGA" }, "expected_output": { "resourceType": "Observation", "status": "final", "code": { "coding": [ { "system": "https://www.ncbi.nlm.nih.gov/gene", "code": "672", "display": "BRCA1" } ] }, "valueQuantity": { "value": 435, "unit": "base pairs" }, "valid_open_reading_frame": true } }, { "input": { "gene_id": "672", "fasta_header": "NM_000000.1 Malformed", "sequence": "ATGGCCT" }, "expected_output": { "resourceType": "Observation", "status": "preliminary", "code": { "coding": [ { "system": "https://www.ncbi.nlm.nih.gov/gene", "code": "672", "display": "BRCA1" } ] }, "valueQuantity": { "value": 7, "unit": "base pairs" }, "valid_open_reading_frame": false } } ] ```

Research-only sequence curation tool. Contains simulated metadata without patient identifiers or diagnostic assertions. Output examples are simplified synthetic app records, not complete official schema files. Use mock external writes. Require human review before any real deployment.