Dataset for: Accelerating multi-objective VHH discovery via integrated high-throughput selection and AlphaFold3-guided structure prediction

structures.tar.xz contains AlphaFold3 predictions output files for all predicted complexes. combined_data.xlsx contains four sheets: all VHH sequences obtained from deep sequencing, annotated with the corresponding yeast display population; all VHH s for which AlphaFold 3 complex predictions were generated with one or more toxins. For each VHH-toxin pair, the table reports ipTM and pTM scores from the top-ranked model; entries are NA where a given VHH was not predicted against that toxin; all VHH s tested in the split-luciferase assay, including ipTM values for each target (αCbtx, αBgtx, cLNTx) and both raw and processed expression and binding measurements; split-luciferase binding and expression data for CamSol- and structural-evolution designed VHH variants, together with their corresponding wild-type sequences. code.zip contains the analysis code used to produce the figures in the manuscript: - notebooks/ — one Jupyter notebook per pipeline stage / figure group; - SRC/ — standalone CLI script for AF3 output preprocessing; - results/ — the input/output data each notebook reads/writes, so notebooks run end-to-end.