PMechRP: Polar Mechanistic Reaction Predictor

Polar Predictor is a deep learning based software that predicts the outcome of polar reactions. PMechRP offers three interfaces: (1) Single Step Predictor, (2) Pathway Predictor, and (3) ArrowFinder. In the single step prediction, the trained deep learning models identify the reactive molecular orbitals and rank their mechanistic interactions based on their chemical plausibility. In the pathway search, the trained deep learning models produce the mechanistic pathways starting from a given set of reactants. The user can search for any desired target within the produced pathways. In ArrowFinder, the model takes in both reactants and products, and predicts the arrow-pushing mechanism.