Schematic diagram of the prediction process for the turnover numbers of enzymatic reactions (IMAGE)
Caption
Enzymes are amino acid sequences; these sequences are converted into numerical vectors, depicted as grey squares, which are then transformed into a single vector by a deep learning model (top left). Information about the catalysed reactions is also converted into numerical vectors (top right). Experimentally determined turnover numbers (bottom left) are used to train a gradient boosting model in order to predict the kcat turnover number (bottom right). Gradient boosting models are an ensemble of multiple decision trees, depicted in different green tones.
Credit
HHU/Alexander Kroll
Usage Restrictions
No restrictions.
License
Public Domain