PesticideDB is a curated database and annotation platform for pesticide-degrading microorganisms, genes, proteins, enzymes, and biodegradation pathways. The resource integrates literature-derived biodegradation evidence with searchable microorganism and protein records, curated pathway information, protein structure assets, and DIAMOND/HMMER-based annotation tools for uploaded gene, protein, genome, or contig FASTA files.

PesticideDB was developed to support research on microbial pesticide biodegradation, environmental bioremediation, and discovery of degradation-associated genes and enzymes. The database provides a structured web interface for browsing pesticide-microorganism associations, protein and enzyme evidence, predicted protein structures, curated transformation pathways, and summary statistics describing taxonomic, biochemical, and literature evidence patterns.

• Curated pesticide biodegradation records linked to microorganisms, isolation sources, publications, genes, enzymes, metabolites, and evidence summaries.
• Searchable microorganism and protein pages for exploring reported pesticide-degrading taxa and degradation-associated proteins.
• Curated pathway and compound pages for reviewing pesticide transformation routes and supporting evidence.
• Interactive annotation tools using DIAMOND and HMMER to compare user-submitted FASTA sequences against PesticideDB reference protein datasets.
• Statistics and visualization pages summarizing database coverage, taxonomic breadth, evidence support, transformation families, and reporting gaps.
• Deployment-ready Django application with local SQLite support and PostgreSQL configuration for shared or production hosting.

• base/: Django application logic, models, views, templates, tests, and management commands.
• PepDatabase/: Django project settings and URL configuration.
• data_files/: Curated core, protein, and pathway input datasets.
• PBDB_annotation/: Reference FASTA files, DIAMOND databases, HMMER profiles, and annotation helper scripts.
• static/: Website styling, images, JavaScript assets, and database summary figures.
• media/protein_structures/: Local predicted protein structure files and preview images.
• curation_outputs/: Supporting curated pathway, taxonomy, evidence, and supplemental protein data exports.

python -m venv env
source env/bin/activate
pip install -r requirements.txt
export DJANGO_DEBUG=True
python manage.py migrate
python manage.py runserver

The web annotation tools call external command-line programs:

• diamond
• hmmscan from HMMER
• prodigal for nucleotide FASTA translation before blastp

Make sure these commands are installed and available on PATH for the Django process.

The annotation assets are read from PBDB_annotation/ by default. To place them elsewhere, set:

export PBDB_ANNOTATION_DIR=/path/to/PBDB_annotation

The app reads these optional environment variables:

• DJANGO_SECRET_KEY
• DJANGO_DEBUG
• DJANGO_ALLOWED_HOSTS
• DJANGO_PUBLIC_DEMO
• DJANGO_CSRF_TRUSTED_ORIGINS
• DJANGO_ENABLE_ADMIN
• DJANGO_SERVE_MEDIA
• DJANGO_ALLOW_DEVTUNNELS
• DATABASE_URL
• DATABASE_SSLMODE
• DATABASE_CONN_MAX_AGE
• PBDB_ANNOTATION_DIR

Annotation jobs write per-job intermediate and downloadable files under MEDIA_ROOT, which prevents simultaneous jobs from overwriting each other.

For broad candidate discovery, the annotation forms default to DIAMOND 1e-3 and 25 percent identity. Users can tighten these thresholds for higher-confidence homologs.

For private multi-user testing, use PostgreSQL instead of SQLite and keep Django in non-debug mode:

export DJANGO_SECRET_KEY=<long-random-secret>
export DJANGO_DEBUG=False
export DJANGO_ALLOWED_HOSTS=<your-private-testing-domain>
export DJANGO_CSRF_TRUSTED_ORIGINS=https://<your-private-testing-domain>
export DJANGO_SECURE_COOKIES=True
export DJANGO_SECURE_SSL_REDIRECT=True
export DJANGO_SECURE_HSTS_SECONDS=31536000
export DJANGO_TRUST_PROXY_SSL_HEADER=True
export DJANGO_ENABLE_ADMIN=False
export DJANGO_SERVE_MEDIA=False
export DATABASE_URL=postgres://USER:PASSWORD@HOST:5432/DBNAME
python manage.py migrate

SQLite remains available only when DATABASE_URL is not set, which is useful for local development but not recommended for several testers using the site at the same time.

For a short VS Code Port Forwarding, ngrok, or Cloudflare Tunnel demo, run Django in public-demo mode. This forces DEBUG=False, disables the Django admin URL, and stops serving raw files from /media/.

After your tunnel gives you a public hostname, run:

scripts/run_public_demo.sh your-public-hostname 8000

Use only the hostname, without https://. For example:

scripts/run_public_demo.sh abc-8000.app.github.dev 8000

Stop the server when the demo is finished. If you need admin access, do it locally in a separate private session rather than through a public tunnel.