Mass spectrometry-based authentication of PDO "Wachauer Marille" using DD-SIMCA one-class classification.

The Wachauer Marille, a protected designation of origin (PDO) apricot from Austria's Wachau region, is prized for quality and flavour, but mislabelling of non-PDO fruit drives the need for reliable authentication. We developed an untargeted metabolomics workflow using LC-QTOFMS fingerprinting and data-driven soft independent modelling of class analogy (DD-SIMCA) to authenticate genuine PDO Wachauer Marille against apricots grown elsewhere in Austria and neighbouring countries. Unlike discriminant models, one-class DD-SIMCA models only the authentic class and can reject non-conforming samples without training on non-PDO alternatives. One-class DD-SIMCA was used solely for authentication, whereas orthogonal partial least squares discriminant analysis (OPLS-DA) served for exploratory discrimination and marker selection. The full 1127-feature model achieved 97% cross-validated sensitivity and correctly rejected 73-100% of non-PDO challenge samples. An exploratory reduced 30-marker model retained 97% cross-validated sensitivity with 60-100% challenge-set rejection. The workflow provides proof-of-concept support for PDO/PGI origin verification.