Prediction of Log P with Property‐Based Methods

Authors: Tetko IV, Poda GI
Publication: Molecular Drug Properties: Measurement and Prediction
Software: ADMET Predictor®


This chapter contains sections titled:

  • Introduction
  • Methods Based on 3D Structure Representation
    • Empirical Approaches
      • LSER
      • SLIPPER
      • SPARC
    • Methods Based on Quantum Chemical Semiempirical Calculations
      • Correlation of Log P with Calculated Quantum Chemical Parameters
      • QLOGP: Importance of Molecular Size
    • Approaches Based on Continuum Solvation Models
      • GBLOGP
      • COSMO‐RS (Full) Approach
      • COSMOfrag (Fragment‐based) Approach
      • Ab Initio Methods
      • QuantlogP
    • Models Based on MD Calculations
    • MLP Methods
      • Early Methods of MLP Calculations
      • Hydrophobic Interactions (HINT)
      • Calculated Lipophilicity Potential (CLIP)
    • Log P Prediction Using Lattice Energies
  • Methods Based on Topological Descriptors
    • MLOGP
    • Graph Molecular Connectivity
      • TLOGP
    • Methods Based on Electrotopological State (E‐state) Descriptors
      • VLOGP
      • ALOGPS
      • CSlogP
      • A_S+logP
  • Prediction Power of Property‐based Approaches
    • Datasets Quality and Consistence
    • Background Models
    • Benchmarking Results
    • Pitfalls of the Benchmarking
      • Do We Compare Methods or Their Implementations?
      • Overlap in the Training and Benchmarking Sets
      • Zwitterions
      • Tautomers and Aromaticity
  • Conclusions