FOX Gene Complement Explorer

Open the live Streamlit app

FOX Gene Complement Explorer turns the published pFOX candidate rankings into a reviewable design tool. Users can search and filter genes, compare conservation evidence across cyanobacteria, and assemble candidate accessory-gene complements under a coding-sequence length budget. It is intended to support experimental planning, not to replace genetic validation.

Paper: Predicting FOX gene candidates for oxic nitrogen fixation using multi-omic machine learning and comparative bioinformatics, Scientific Reports 16, 11412 (2026)

Modeling code: FoxGenes_ML

Key result

The associated paper reports that the combined multi-omic models ranked held-out FOX genes above the 0.075 prevalence baseline and released genome-wide pFOX scores for candidate prioritization. This app applies the paper's comparative-bioinformatics filters and size constraints to those rankings. pFOX is not a calibrated probability of FOX essentiality.

What the app provides

• Gene search and model-score filtering
• Filamentous and unicellular diazotroph conservation filters
• Non-diazotroph exclusion filters
• Rank-based and probability-per-length complement selection
• Ortholog panels, summary statistics, word clouds, and CSV downloads

Install and run

git clone https://github.com/jamesyoung93/FoxGenesApp.git
cd FoxGenesApp
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
streamlit run app.py

On Windows PowerShell, activate the environment with .venv\Scripts\Activate.ps1.

Stack

Python, Streamlit, pandas, matplotlib, matplotlib-venn, and wordcloud.

Data availability

The repository includes the candidate table loaded by the app. The table is derived from the public modeling workflow in FoxGenes_ML. The underlying public data sources and accession identifiers are documented in the paper's Data Availability section.

Citation

@article{Young2026FoxGenes,
  author  = {Young, James and Gu, Liping and Zhou, Ruanbao},
  title   = {Predicting FOX gene candidates for oxic nitrogen fixation using multi-omic machine learning and comparative bioinformatics},
  journal = {Scientific Reports},
  year    = {2026},
  volume  = {16},
  pages   = {11412},
  doi     = {10.1038/s41598-026-41873-w}
}

License

Code is available under the MIT License. The bundled candidate table and third-party source data retain their original terms.