Phenotype

Phenotype is a local genome annotation app for inspecting SNPs without pushing raw genotype data to a hosted service. It combines imported genotypes with cached sources such as ClinVar, SNPedia, VEP, and Promethease-style reports.

What It Does

• Upload a 23andMe or Ancestry-style genome file.
• Review Findings, New, Clinical, All annotated, Unannotated, and Build 37 variants.
• Search by rsid or gene.
• Filter by zygosity, VEP impact, VEP consequence, and finding severity.
• Sort by magnitude, rating, recency, publication count, gene, rsid, and frequency.
• Open a SNP in the sidebar to see linked studies, SNPedia, ClinVar, and metadata.
• Import VEP output, ClinVar data, and Promethease HTML reports.
• Export VEP input for build 37 variants or rsid lists.

Why It Feels Fast

• The table is summary-first.
• Full SNP detail loads only when you click a row.
• Summary counts and the default findings view are cached locally.
• Matching is stricter: rows only show as findings when the imported genotype actually matches.

Screenshot

Quick Start

python3 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install -e ".[dev]"

Run the app:

cd src/phenotype-web
../../.venv/bin/python -m flask --app phenotype.app:create_app run --host 127.0.0.1 --port 5000

Open the local URL printed by Flask. If 5000 is busy, use another free port.

Local data lives under src/phenotype-web/data/:

• scrapedData.json for cached annotations
• yourData.json for imported genotypes
• phenotype.sqlite for the local cache

Importing Data

Genome file

Use the browser upload or run the importer directly:

cd src/phenotype-web
../../.venv/bin/python -m phenotype.scraper

Promethease HTML report

Import a Promethease-style HTML report from the browser to seed the cache with existing findings.

ClinVar and VEP

The browser supports:

• Import ClinVar DB
• Scan x;y ClinVar matches
• Import VEP
• Import report HTML
• Refresh finding dates
• Refresh annotations

VEP Workflow

After importing a build 37 genome file, use Build 37 variants with x;y to narrow the export to heterozygous rows.

Install the GRCh37 cache:

mkdir -p "$HOME/vep_data"
docker run -t -i -v "$HOME/vep_data:/data" ensemblorg/ensembl-vep INSTALL.pl -a cf -s homo_sapiens -y GRCh37

Run VEP on the exported build 37 variants:

docker run --rm \
  -v "$HOME/vep_data:/data" \
  -v "$PWD/src/phenotype-web/data/exports:/work" \
  ensemblorg/ensembl-vep \
  vep --cache --offline --assembly GRCh37 --format ensembl \
  --input_file /work/phenotype_build37_heterozygous_vep_input.tsv \
  --output_file /work/phenotype_vep_output.txt \
  --force_overwrite --tab --symbol --hgvs --canonical --variant_class --no_stats

Import the result:

.venv/bin/phenotype-vep-import src/phenotype-web/data/exports/phenotype_vep_output.txt

API

• GET /api/snps
• GET /api/snps/<rsid>
• GET /api/variants
• GET /api/backlog
• POST /api/import
• POST /api/report/import
• POST /api/scrape
• POST /api/scrape/resume
• POST /api/refresh-findings
• POST /api/vep/import
• GET /api/scrape-runs/latest
• GET /api/scrape-runs/<run_id>
• GET /api/export.csv
• GET /api/export-vep.tsv
• GET /api/export-vep-rsids.txt
• DELETE /api/genotypes

Development

.venv/bin/python -m pytest
.venv/bin/python -m ruff check src/phenotype-web/phenotype tests
.venv/bin/python -m ruff format src/phenotype-web/phenotype tests

Notes

• The app is for personal exploration, not diagnosis.
• Direct-to-consumer raw data can contain false positives.
• Clinical interpretation should be confirmed with an appropriate lab test and a clinician.