Data for: LEGO-based Generalized Set of Two Linear Algebraic 3D Bio-Macro-Molecular Descriptors: Theory and Validation by QSARs
Description
SI3-1: 15 suggested theoretical configurations for the calculation of MDs (defined with the name projects). The selected configuration for the projects used in this study are also indicated in Table SI2-1 and are available at SI3-1. SI3-2: The experiments employed a dataset containing 152 representatives, non-homologous proteins (see SI3-2 to review the protein files). (Fleming and Richards, 2000). SI3-3: The evaluation of this application in protein science requires the use of two datasets. The first data set, employed as a training set, was proposed by Ouyang (Ouyang and Liang, 2008) and contains 80 proteins (the case “2BLM” was removed since it only considered an alpha carbon representation). The second dataset, employed as a test set, was proposed by Ruiz-Blanco (Ruiz-Blanco et al., 2015) and contains 17 proteins. SI3-4: For the generation of the models, a dataset of 204 proteins was employed. (Chou, 1999) The original dataset was split into two groups: a training set with 149 proteins and a test set containing 55 proteins to ensure a proper comparison (Marrero Ponce et al., 2015a)