SAMPL6: calculation of macroscopic pK a values from ab initio quantum mechanical free energies

Edithe Selwa, Ian M. Kenney, Oliver Beckstein, Bogdan I. Iorga

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Macroscopic pKa values were calculated for all compounds in the SAMPL6 blind prediction challenge, based on quantum chemical calculations with a continuum solvation model and a linear correction derived from a small training set. Microscopic pKa values were derived from the gas-phase free energy difference between protonated and deprotonated forms together with the Conductor-like Polarizable Continuum Solvation Model and the experimental solvation free energy of the proton. pH-dependent microstate free energies were obtained from the microscopic pKas with a maximum likelihood estimator and appropriately summed to yield macroscopic pKa values or microstate populations as function of pH. We assessed the accuracy of three approaches to calculate the microscopic pKas: direct use of the quantum mechanical free energy differences and correction of the direct values for short-comings in the QM solvation model with two different linear models that we independently derived from a small training set of 38 compounds with known pKa. The predictions that were corrected with the linear models had much better accuracy [root-mean-square error (RMSE) 2.04 and 1.95 pKa units] than the direct calculation (RMSE 3.74). Statistical measures indicate that some systematic errors remain, likely due to differences in the SAMPL6 data set and the small training set with respect to their interactions with water. Overall, the current approach provides a viable physics-based route to estimate macroscopic pKa values for novel compounds with reasonable accuracy.

Original languageEnglish (US)
JournalJournal of Computer-Aided Molecular Design
DOIs
StateAccepted/In press - Jan 1 2018

Fingerprint

Free energy
Solvation
Linear Models
free energy
solvation
education
root-mean-square errors
Mean square error
Physics
Protons
continuums
Theoretical Models
Gases
Systematic errors
predictions
Water
estimators
systematic errors
Maximum likelihood
conductors

Keywords

  • pH
  • pK
  • Quantum chemistry
  • SAMPL challenge

ASJC Scopus subject areas

  • Drug Discovery
  • Computer Science Applications
  • Physical and Theoretical Chemistry

Cite this

SAMPL6 : calculation of macroscopic pK a values from ab initio quantum mechanical free energies. / Selwa, Edithe; Kenney, Ian M.; Beckstein, Oliver; Iorga, Bogdan I.

In: Journal of Computer-Aided Molecular Design, 01.01.2018.

Research output: Contribution to journalArticle

@article{a1f4b2a7ae064154a87bbd734b698e77,
title = "SAMPL6: calculation of macroscopic pK a values from ab initio quantum mechanical free energies",
abstract = "Macroscopic pKa values were calculated for all compounds in the SAMPL6 blind prediction challenge, based on quantum chemical calculations with a continuum solvation model and a linear correction derived from a small training set. Microscopic pKa values were derived from the gas-phase free energy difference between protonated and deprotonated forms together with the Conductor-like Polarizable Continuum Solvation Model and the experimental solvation free energy of the proton. pH-dependent microstate free energies were obtained from the microscopic pKas with a maximum likelihood estimator and appropriately summed to yield macroscopic pKa values or microstate populations as function of pH. We assessed the accuracy of three approaches to calculate the microscopic pKas: direct use of the quantum mechanical free energy differences and correction of the direct values for short-comings in the QM solvation model with two different linear models that we independently derived from a small training set of 38 compounds with known pKa. The predictions that were corrected with the linear models had much better accuracy [root-mean-square error (RMSE) 2.04 and 1.95 pKa units] than the direct calculation (RMSE 3.74). Statistical measures indicate that some systematic errors remain, likely due to differences in the SAMPL6 data set and the small training set with respect to their interactions with water. Overall, the current approach provides a viable physics-based route to estimate macroscopic pKa values for novel compounds with reasonable accuracy.",
keywords = "pH, pK, Quantum chemistry, SAMPL challenge",
author = "Edithe Selwa and Kenney, {Ian M.} and Oliver Beckstein and Iorga, {Bogdan I.}",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/s10822-018-0138-6",
language = "English (US)",
journal = "Journal of Computer-Aided Molecular Design",
issn = "0920-654X",
publisher = "Springer Netherlands",

}

TY - JOUR

T1 - SAMPL6

T2 - calculation of macroscopic pK a values from ab initio quantum mechanical free energies

AU - Selwa, Edithe

AU - Kenney, Ian M.

AU - Beckstein, Oliver

AU - Iorga, Bogdan I.

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Macroscopic pKa values were calculated for all compounds in the SAMPL6 blind prediction challenge, based on quantum chemical calculations with a continuum solvation model and a linear correction derived from a small training set. Microscopic pKa values were derived from the gas-phase free energy difference between protonated and deprotonated forms together with the Conductor-like Polarizable Continuum Solvation Model and the experimental solvation free energy of the proton. pH-dependent microstate free energies were obtained from the microscopic pKas with a maximum likelihood estimator and appropriately summed to yield macroscopic pKa values or microstate populations as function of pH. We assessed the accuracy of three approaches to calculate the microscopic pKas: direct use of the quantum mechanical free energy differences and correction of the direct values for short-comings in the QM solvation model with two different linear models that we independently derived from a small training set of 38 compounds with known pKa. The predictions that were corrected with the linear models had much better accuracy [root-mean-square error (RMSE) 2.04 and 1.95 pKa units] than the direct calculation (RMSE 3.74). Statistical measures indicate that some systematic errors remain, likely due to differences in the SAMPL6 data set and the small training set with respect to their interactions with water. Overall, the current approach provides a viable physics-based route to estimate macroscopic pKa values for novel compounds with reasonable accuracy.

AB - Macroscopic pKa values were calculated for all compounds in the SAMPL6 blind prediction challenge, based on quantum chemical calculations with a continuum solvation model and a linear correction derived from a small training set. Microscopic pKa values were derived from the gas-phase free energy difference between protonated and deprotonated forms together with the Conductor-like Polarizable Continuum Solvation Model and the experimental solvation free energy of the proton. pH-dependent microstate free energies were obtained from the microscopic pKas with a maximum likelihood estimator and appropriately summed to yield macroscopic pKa values or microstate populations as function of pH. We assessed the accuracy of three approaches to calculate the microscopic pKas: direct use of the quantum mechanical free energy differences and correction of the direct values for short-comings in the QM solvation model with two different linear models that we independently derived from a small training set of 38 compounds with known pKa. The predictions that were corrected with the linear models had much better accuracy [root-mean-square error (RMSE) 2.04 and 1.95 pKa units] than the direct calculation (RMSE 3.74). Statistical measures indicate that some systematic errors remain, likely due to differences in the SAMPL6 data set and the small training set with respect to their interactions with water. Overall, the current approach provides a viable physics-based route to estimate macroscopic pKa values for novel compounds with reasonable accuracy.

KW - pH

KW - pK

KW - Quantum chemistry

KW - SAMPL challenge

UR - http://www.scopus.com/inward/record.url?scp=85051703600&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051703600&partnerID=8YFLogxK

U2 - 10.1007/s10822-018-0138-6

DO - 10.1007/s10822-018-0138-6

M3 - Article

C2 - 30084080

AN - SCOPUS:85051703600

JO - Journal of Computer-Aided Molecular Design

JF - Journal of Computer-Aided Molecular Design

SN - 0920-654X

ER -