Parkinson protein classification dataset (cleaned version)

This version contains a cleaned release of the Parkinson protein classification dataset. Changes introduced in this version:- Removal of one duplicated protein sequence appearing in both classes, which represented a potential source of data leakage. One sequence containing a non-standard amino acid symbol (U, selenocysteine) was identified but retained in this version of the dataset. The resulting curated dataset contains 304 protein sequences. This dataset was constructed from UniProt records and is intended for binary classification tasks involving Parkinson-associated and control proteins.