作者: Andrew G Mercader , Pablo R Duchowicz , Miguel A Sanservino , Francisco M Fernández , Eduardo A Castro
DOI: 10.1016/J.JFLUCHEM.2006.12.011
关键词: Chemistry 、 Quantitative structure–activity relationship 、 Computational chemistry 、 Linear regression 、 Replacement method 、 Molecular descriptor 、 Correlation coefficient 、 Molecule
摘要: Abstract We constructed a QSPR model from 116 organic compounds for the prediction of fluorophilicity. The 1268 theoretical descriptors explored by means linear regressions, encoding different aspects topological, geometrical, and electronic molecular structure, lead to an optimal seven-parameter equation with correlation coefficient R = 0.9807 cross-validation parameter Rl-15%-o = 0.9677. As more realistic practical application present model, it is applied estimation fluorophilicity 69 non-yet synthesized structures.