itamar-menuhin/STABLES_code: STABLES v1.0

🛠️ Release: STABLES v1.0 This release accompanies the publication:"AI-directed gene fusing prolongs the evolutionary half-life of synthetic gene circuits"by Menuhin-Gruman et al., Science Advances (2025) 📦 What's included: Full source code for feature engineering, model training, and fusion prediction Analysis scripts for linker and stop codon selection Datasets used in the computational pipeline and validation experiments All figures used in the main and supplementary material Environment files (requirements.txt) for full reproducibility 🔬 Highlights: Predictive ensemble model (XGBoost + KNN) for selecting endogenous genes (EGs) to stabilize synthetic constructs Pipeline to design leaky stop codon fusions with enhanced expression stability Extensive experimental validation in S. cerevisiae, including proinsulin expression Feature importance via SHAP analysis 🧬 Data Use Notes: Yeast genome annotations were derived from publicly available resources (SGD, Ensembl) and processed for feature extraction. Due to size and open accessibility, raw genome files are not hosted here. Scripts and processed summaries (genome.xlsx) are included. 📄 License: This work is released under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license. You are free to use, share, and adapt this material for non-commercial purposes, provided appropriate credit is given. 🔗 Citation: When using this code or data, please cite both the paper and this Zenodo release: Menuhin-Gruman, I. et al. (2025). AI-directed gene fusing prolongs the evolutionary half-life of synthetic gene circuits. Science Advances.Zenodo DOI: 10.5281/zenodo.16959725