This repository contains the custom computational pipelines and analysis scripts required to reproduce the in silico findings presented in the manuscript "A Zero-Shot Generative Framework for Stable Genetic Circuits to Model Chemotherapy Toxicity." The codebase utilizes the Nucleotide Transformer (a genomic foundation model) to perform zero-shot inference on synthetic promoter sequences. It includes scripts for: Calculating evolutionary fitness scores and mapping the biophysical expression-fitness trade-off. Generating 2D KDE Topographical Quadrant Maps to identify mutation-resistant "Optimal Trade-off" sequences. Performing in silico saturation mutagenesis on specific promoter architectures. Executing computationally directed evolution to upgrade candidate sequences prior to physical synthesis. Requirements: Python 3.9+, PyTorch, HuggingFace Transformers, Pandas, and Seaborn. See README.md for full installation and execution instructions.