Data and model files for FCKcat: prediction of enzyme kcat fold-changes

This record contains the data and model files required to reproduce the analyses and figures in the accompanying GitHub repository for the manuscript “Overcoming systematic data biases enables accurate prediction of enzyme kcat fold-changes for computational protein design”.

The accompanying code and notebooks are available at: https://github.com/yvanrousset/FCKcat

The archive contains the data/ directory expected by the notebooks in the FCKcat GitHub repository. It includes train/test pair dataframes, the trained FCKcat XGBoost model, ESM2 embeddings and sequence mapping files, baseline prediction files and metadata for DLKcat and EITLEM comparisons, and aggregated evaluation results.

To use this archive, download `data.zip` and extract it into the root of the FCKcat GitHub repository. This should create a `data/` folder containing the required subdirectories:

datasets/
models/
mapping/
EITLEM_data/
DLKcat_data/
results/

These files are intended to accompany the public code repository and enable reproduction of the analyses reported in the manuscript.