Skip to content

cli.md

Latest commit

 

History

History
64 lines (45 loc) · 4.12 KB

cli.md

File metadata and controls

64 lines (45 loc) · 4.12 KB
Raw

Command-line options

On the very first run, gnomancer FILE downloads the open-source tools and reference files automatically (same as gnomancer setup); pass --no-setup to prevent that.

gnomancer FILE [--out DIR] [--markers TSV] [--no-y] [--no-yleaf] [--no-mt] [--no-clinvar] [--no-pharmcat] [--ancestry]
          [--summary summary.md] [--ai] [--no-setup]
gnomancer --version

Plain-language summary. The report opens with an automatic "at a glance" block built from the results. For a narrative on top of that, every run also writes summary-prompt.md: paste it to whatever AI assistant you use (Claude, ChatGPT, a local model), save the answer as summary.md, and re-run with --summary summary.md. Or pass --ai to have Claude write it directly (pip install anthropic, credentials via ANTHROPIC_API_KEY or ant auth login). --ai is the one path that sends anything off your machine: the structured results in summary-prompt.md, never the raw genotype file.

--ancestry runs ezancestry for a 1000 Genomes population classification. ezancestry and Yleaf need Python ≤3.11; gnomancer setup builds a side venv (~/.gnomancer/venv-ancestry) for both automatically when python3.11 is on the PATH, and the CLI uses it when present. First runs download reference data (ezancestry models; Yleaf's hg19 reference, ~900 MB).

Try it without your own data

python examples/make_demo.py demo      # synthetic AncestryDNA / 23andMe / MyHeritage files (from a repo checkout)
gnomancer demo/demo_AncestryDNA.txt

Pharmacist card

Every run also writes pgx-card.pdf (and pgx-card.html): a one-page summary of the PharmCAT diplotypes, phenotypes and the drugs each gene matters for, with the clotting and abacavir single-site results and a footer that says it is array-based and provisional. Print it or show it on a phone; the report links to it from the PharmCAT section.

gnomancer family KIT1 KIT2 [KIT3 ...]: compare relatives

Give it two or more raw files (any vendors). It writes family-report.html / .md with:

  • Shared DNA per pair (segments via apriha/lineage on the HapMap genetic map, 7 cM minimum) and a relationship estimate from total centimorgans. Parent/child and full siblings are called confidently; the overlapping classes (grandparent / half-sibling / aunt-uncle, first cousin / great-grandparent, ...) are listed as "one of these".
  • Shared lineages: same mtDNA haplogroup means the same maternal line; same Y means the same paternal line.
  • Panel sites side by side, and for a parent-parent-child set, which parent each of the child's alleles came from where that is determinable.
  • Mendelian-consistency check for trios: well above 1% inconsistencies means a mislabeled kit.

Options: --names dad,mom,me (labels in file order), --out DIR, --no-haplogroups (faster). First run downloads the HapMap map (~20 MB). Segment plots per pair land in shared-dna/. From an assistant: the MCP tool family.

gnomancer watch REPORT_DIR: ClinVar reclassification check

ClinVar reclassifies variants every month. watch keeps a snapshot of every ClinVar-classified allele you carry (any significance) next to your report (clinvar-snapshot.json, written on every normal run), checks whether NCBI has published a newer ClinVar release, downloads it if so, rescans the same raw file, and prints only what changed:

  • newly pathogenic / likely pathogenic at a site you carry (new record or reclassification)
  • downgraded (no longer pathogenic)
  • review-status changes on pathogenic sites

The result is also saved as clinvar-watch.md in the report folder. Flags: --force (rescan even if ClinVar is unchanged), --no-download (use the ClinVar file already on disk). The raw file must still be at the path recorded in summary.json.

Run it monthly. macOS/Linux cron, 9:00 on the 2nd of each month (ClinVar releases on the 1st):

0 9 2 * * ~/.gnomancer/venv/bin/gnomancer watch ~/Downloads/AncestryDNA_analysis >> ~/.gnomancer/watch.log 2>&1

From an AI assistant with the MCP server connected: ask it to run clinvar_watch.