The dataset is constituted by 79 proteins in the training set and 16 in the test set. The purpose of this dataset is to assess the capability in describing the protein structure of the new algebraic form-based 3D-protein descriptors by predicting the folding rate constant.