Novel Methods for the Prediction of logP, pK_a, and logD

Abstract

Novel methods for predicting logP, pK(a), and logD values have been developed using data sets (592 molecules for logP and 1029 for pK(a)) containing a wide range of molecular structures. An equation with three molecular properties (polarizability and partial atomic charges on nitrogen and oxygen) correlates highly with logP (r2 = 0.89). The pK(a)s are estimated for both acids and bases using a novel tree structured fingerprint describing the ionizing centers. The new models have been compared with existing models and also experimental measurements on test sets of common organic compounds and pharmaceutical molecules.

Predicting and Tuning Physicochemical Properties in Lead Optimization: Amine Basicities

“…[40] Based on a compound set of 384 basic and 645 acidic aqueous pK a values extracted from Lange's Handbook of Chemistry, [41] an ionization model was generated. Combined descriptors ranked by cross-validated partial leastsquares analysis were mapped onto molecular tree constructs around the ionizable center.…”

Computational Methods in Drug Discovery

“…Additionally, correction factors are included for aromatic nitrogen pairs, ortho sp 3 oxygen pairs, para donor pairs, sp 2 oxygen pairs, and amino sulfonic acids. Xing and Glen (2002) introduced an alternative logP calculation that was based on the evidence that molecular size and hydrogen-bonding ability account for a major part of logP. They created a statistical model by combining molecular size and dispersion interactions using molecular polarizability and the sum of squared partial atomic charges on oxygen and nitrogen atoms.…”

Analysis of Peptide−Protein Binding Using Amino Acid Descriptors: Prediction and Experimental Verification for Human Histocompatibility Complex HLA-A*0201

“…Many of these properties are measured experimentally by methods such as TLC, HPLC, and spectroscopy. [14][15][16] Other properties cannot be measured but can be calculated, such as molecular surface areas and atomic charges. The quality of the amino acid descriptors is an important factor in producing models with good predictivity.…”