First stable release of AAanalysis (Amino Acid analysis), a Python framework for interpretable sequence-based protein prediction.
This version includes the foundational algorithms used in the publication "Charting γ-secretase substrates by explainable AI" (Breimann & Kamp et al., Nature Communications, 2025):
CPP (Comparative Physicochemical Profiling), a feature engineering method that identifies the most distinctive physicochemical properties between two sets of protein sequences, and dPULearn, a deterministic positive-unlabeled (PU) learning algorithm enabling robust classification from imbalanced and small datasets.