AlphaFold DB Thermophile Protein Structures (Global Confidence 90–100)

This dataset contains high-confidence protein structure predictions for thermophile-annotated proteins retrieved from the AlphaFold Protein Structure Database. Entries were identified using a text-based search for “Thermophile” and filtered to include only models with global confidence scores between 90 and 100. For each protein entry, the dataset provides AlphaFold-predicted atomic coordinates in PDB format, Predicted Aligned Error (PAE) matrices in JSON format, and the corresponding amino acid sequences in FASTA format. Only entries for which both structure coordinates and PAE data were available were included. The dataset was generated using a custom Python script that queries the AlphaFold DB API and downloads the associated files in a reproducible manner. This resource is intended for use in structural bioinformatics, protein stability analysis, and machine-learning applications involving thermophilic proteins.