Application of a single-objective, hybrid genetic algorithm approach to pharmacokinetic model building

Eric A. Sherer, Mark E. Sale, Bruce G. Pollock, Chandra P. Belani, Merrill J. Egorin, Percy S. Ivy, Jeffrey A. Lieberman, Stephen B. Manuck, Stephen R. Marder, Matthew F. Muldoon, Howard I. Scher, David B. Solit, Robert Bies

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

A limitation in traditional stepwise population pharmacokinetic model building is the difficulty in handling interactions between model components. To address this issue, a method was previously introduced which couples NONMEM parameter estimation and model fitness evaluation to a single-objective, hybrid genetic algorithm for global optimization of the model structure. In this study, the generalizability of this approach for pharmacokinetic model building is evaluated by comparing (1) correct and spurious covariate relationships in a simulated dataset resulting from automated stepwise covariate modeling, Lasso methods, and single-objective hybrid genetic algorithm approaches to covariate identification and (2) information criteria values, model structures, convergence, and model parameter values resulting from manual stepwise versus single-objective, hybrid genetic algorithm approaches to model building for seven compounds. Both manual stepwise and single-objective, hybrid genetic algorithm approaches to model building were applied, blinded to the results of the other approach, for selection of the compartment structure as well as inclusion and model form of inter-individual and inter-occasion variability, residual error, and covariates from a common set of model options. For the simulated dataset, stepwise covariate modeling identified three of four true covariates and two spurious covariates; Lasso identified two of four true and 0 spurious covariates; and the single-objective, hybrid genetic algorithm identified three of four true covariates and one spurious covariate. For the clinical datasets, the Akaike information criterion was a median of 22.3 points lower (range of 470.5 point decrease to 0.1 point decrease) for the best single-objective hybrid genetic-algorithm candidate model versus the final manual stepwise model: the Akaike information criterion was lower by greater than 10 points for four compounds and differed by less than 10 points for three compounds. The root mean squared error and absolute mean prediction error of the best singleobjective hybrid genetic algorithm candidates were a median of 0.2 points higher (range of 38.9 point decrease to 27.3 point increase) and 0.02 points lower (range of 0.98 point decrease to 0.74 point increase), respectively, than that of the final stepwise models. In addition, the best single-objective, hybrid genetic algorithm candidate models had successful convergence and covariance steps for each compound, used the same compartment structure as the manual stepwise approach for 6 of 7 (86 %) compounds, and identified 54 % (7 of 13) of covariates included by the manual stepwise approach and 16 covariate relationships not included by manual stepwise models. The model parameter values between the final manual stepwise and best single-objective, hybrid genetic algorithm models differed by a median of 26.7 % (q1 = 4.9 % and q3 = 57.1 %). Finally, the single-objective, hybrid genetic algorithm approach was able to identify models capable of estimating absorption rate parameters for four compounds that the manual stepwise approach did not identify. The single-objective, hybrid genetic algorithm represents a general pharmacokinetic model building methodology whose ability to rapidly search the feasible solution space leads to nearly equivalent or superior model fits to pharmacokinetic data.

Original languageEnglish
Pages (from-to)393-414
Number of pages22
JournalJournal of Pharmacokinetics and Pharmacodynamics
Volume39
Issue number4
DOIs
StatePublished - Aug 2012

Fingerprint

Pharmacokinetics
Genetic Models
Population
Datasets

Keywords

  • Genetic Algorithms
  • Model building
  • Pharmacokinetics

ASJC Scopus subject areas

  • Pharmacology

Cite this

Application of a single-objective, hybrid genetic algorithm approach to pharmacokinetic model building. / Sherer, Eric A.; Sale, Mark E.; Pollock, Bruce G.; Belani, Chandra P.; Egorin, Merrill J.; Ivy, Percy S.; Lieberman, Jeffrey A.; Manuck, Stephen B.; Marder, Stephen R.; Muldoon, Matthew F.; Scher, Howard I.; Solit, David B.; Bies, Robert.

In: Journal of Pharmacokinetics and Pharmacodynamics, Vol. 39, No. 4, 08.2012, p. 393-414.

Research output: Contribution to journalArticle

Sherer, EA, Sale, ME, Pollock, BG, Belani, CP, Egorin, MJ, Ivy, PS, Lieberman, JA, Manuck, SB, Marder, SR, Muldoon, MF, Scher, HI, Solit, DB & Bies, R 2012, 'Application of a single-objective, hybrid genetic algorithm approach to pharmacokinetic model building', Journal of Pharmacokinetics and Pharmacodynamics, vol. 39, no. 4, pp. 393-414. https://doi.org/10.1007/s10928-012-9258-0
Sherer, Eric A. ; Sale, Mark E. ; Pollock, Bruce G. ; Belani, Chandra P. ; Egorin, Merrill J. ; Ivy, Percy S. ; Lieberman, Jeffrey A. ; Manuck, Stephen B. ; Marder, Stephen R. ; Muldoon, Matthew F. ; Scher, Howard I. ; Solit, David B. ; Bies, Robert. / Application of a single-objective, hybrid genetic algorithm approach to pharmacokinetic model building. In: Journal of Pharmacokinetics and Pharmacodynamics. 2012 ; Vol. 39, No. 4. pp. 393-414.
@article{f9fc4a4138504a3fb70b37bc149174b7,
title = "Application of a single-objective, hybrid genetic algorithm approach to pharmacokinetic model building",
abstract = "A limitation in traditional stepwise population pharmacokinetic model building is the difficulty in handling interactions between model components. To address this issue, a method was previously introduced which couples NONMEM parameter estimation and model fitness evaluation to a single-objective, hybrid genetic algorithm for global optimization of the model structure. In this study, the generalizability of this approach for pharmacokinetic model building is evaluated by comparing (1) correct and spurious covariate relationships in a simulated dataset resulting from automated stepwise covariate modeling, Lasso methods, and single-objective hybrid genetic algorithm approaches to covariate identification and (2) information criteria values, model structures, convergence, and model parameter values resulting from manual stepwise versus single-objective, hybrid genetic algorithm approaches to model building for seven compounds. Both manual stepwise and single-objective, hybrid genetic algorithm approaches to model building were applied, blinded to the results of the other approach, for selection of the compartment structure as well as inclusion and model form of inter-individual and inter-occasion variability, residual error, and covariates from a common set of model options. For the simulated dataset, stepwise covariate modeling identified three of four true covariates and two spurious covariates; Lasso identified two of four true and 0 spurious covariates; and the single-objective, hybrid genetic algorithm identified three of four true covariates and one spurious covariate. For the clinical datasets, the Akaike information criterion was a median of 22.3 points lower (range of 470.5 point decrease to 0.1 point decrease) for the best single-objective hybrid genetic-algorithm candidate model versus the final manual stepwise model: the Akaike information criterion was lower by greater than 10 points for four compounds and differed by less than 10 points for three compounds. The root mean squared error and absolute mean prediction error of the best singleobjective hybrid genetic algorithm candidates were a median of 0.2 points higher (range of 38.9 point decrease to 27.3 point increase) and 0.02 points lower (range of 0.98 point decrease to 0.74 point increase), respectively, than that of the final stepwise models. In addition, the best single-objective, hybrid genetic algorithm candidate models had successful convergence and covariance steps for each compound, used the same compartment structure as the manual stepwise approach for 6 of 7 (86 {\%}) compounds, and identified 54 {\%} (7 of 13) of covariates included by the manual stepwise approach and 16 covariate relationships not included by manual stepwise models. The model parameter values between the final manual stepwise and best single-objective, hybrid genetic algorithm models differed by a median of 26.7 {\%} (q1 = 4.9 {\%} and q3 = 57.1 {\%}). Finally, the single-objective, hybrid genetic algorithm approach was able to identify models capable of estimating absorption rate parameters for four compounds that the manual stepwise approach did not identify. The single-objective, hybrid genetic algorithm represents a general pharmacokinetic model building methodology whose ability to rapidly search the feasible solution space leads to nearly equivalent or superior model fits to pharmacokinetic data.",
keywords = "Genetic Algorithms, Model building, Pharmacokinetics",
author = "Sherer, {Eric A.} and Sale, {Mark E.} and Pollock, {Bruce G.} and Belani, {Chandra P.} and Egorin, {Merrill J.} and Ivy, {Percy S.} and Lieberman, {Jeffrey A.} and Manuck, {Stephen B.} and Marder, {Stephen R.} and Muldoon, {Matthew F.} and Scher, {Howard I.} and Solit, {David B.} and Robert Bies",
year = "2012",
month = "8",
doi = "10.1007/s10928-012-9258-0",
language = "English",
volume = "39",
pages = "393--414",
journal = "Journal of Pharmacokinetics and Pharmacodynamics",
issn = "1567-567X",
publisher = "Springer New York",
number = "4",

}

TY - JOUR

T1 - Application of a single-objective, hybrid genetic algorithm approach to pharmacokinetic model building

AU - Sherer, Eric A.

AU - Sale, Mark E.

AU - Pollock, Bruce G.

AU - Belani, Chandra P.

AU - Egorin, Merrill J.

AU - Ivy, Percy S.

AU - Lieberman, Jeffrey A.

AU - Manuck, Stephen B.

AU - Marder, Stephen R.

AU - Muldoon, Matthew F.

AU - Scher, Howard I.

AU - Solit, David B.

AU - Bies, Robert

PY - 2012/8

Y1 - 2012/8

N2 - A limitation in traditional stepwise population pharmacokinetic model building is the difficulty in handling interactions between model components. To address this issue, a method was previously introduced which couples NONMEM parameter estimation and model fitness evaluation to a single-objective, hybrid genetic algorithm for global optimization of the model structure. In this study, the generalizability of this approach for pharmacokinetic model building is evaluated by comparing (1) correct and spurious covariate relationships in a simulated dataset resulting from automated stepwise covariate modeling, Lasso methods, and single-objective hybrid genetic algorithm approaches to covariate identification and (2) information criteria values, model structures, convergence, and model parameter values resulting from manual stepwise versus single-objective, hybrid genetic algorithm approaches to model building for seven compounds. Both manual stepwise and single-objective, hybrid genetic algorithm approaches to model building were applied, blinded to the results of the other approach, for selection of the compartment structure as well as inclusion and model form of inter-individual and inter-occasion variability, residual error, and covariates from a common set of model options. For the simulated dataset, stepwise covariate modeling identified three of four true covariates and two spurious covariates; Lasso identified two of four true and 0 spurious covariates; and the single-objective, hybrid genetic algorithm identified three of four true covariates and one spurious covariate. For the clinical datasets, the Akaike information criterion was a median of 22.3 points lower (range of 470.5 point decrease to 0.1 point decrease) for the best single-objective hybrid genetic-algorithm candidate model versus the final manual stepwise model: the Akaike information criterion was lower by greater than 10 points for four compounds and differed by less than 10 points for three compounds. The root mean squared error and absolute mean prediction error of the best singleobjective hybrid genetic algorithm candidates were a median of 0.2 points higher (range of 38.9 point decrease to 27.3 point increase) and 0.02 points lower (range of 0.98 point decrease to 0.74 point increase), respectively, than that of the final stepwise models. In addition, the best single-objective, hybrid genetic algorithm candidate models had successful convergence and covariance steps for each compound, used the same compartment structure as the manual stepwise approach for 6 of 7 (86 %) compounds, and identified 54 % (7 of 13) of covariates included by the manual stepwise approach and 16 covariate relationships not included by manual stepwise models. The model parameter values between the final manual stepwise and best single-objective, hybrid genetic algorithm models differed by a median of 26.7 % (q1 = 4.9 % and q3 = 57.1 %). Finally, the single-objective, hybrid genetic algorithm approach was able to identify models capable of estimating absorption rate parameters for four compounds that the manual stepwise approach did not identify. The single-objective, hybrid genetic algorithm represents a general pharmacokinetic model building methodology whose ability to rapidly search the feasible solution space leads to nearly equivalent or superior model fits to pharmacokinetic data.

AB - A limitation in traditional stepwise population pharmacokinetic model building is the difficulty in handling interactions between model components. To address this issue, a method was previously introduced which couples NONMEM parameter estimation and model fitness evaluation to a single-objective, hybrid genetic algorithm for global optimization of the model structure. In this study, the generalizability of this approach for pharmacokinetic model building is evaluated by comparing (1) correct and spurious covariate relationships in a simulated dataset resulting from automated stepwise covariate modeling, Lasso methods, and single-objective hybrid genetic algorithm approaches to covariate identification and (2) information criteria values, model structures, convergence, and model parameter values resulting from manual stepwise versus single-objective, hybrid genetic algorithm approaches to model building for seven compounds. Both manual stepwise and single-objective, hybrid genetic algorithm approaches to model building were applied, blinded to the results of the other approach, for selection of the compartment structure as well as inclusion and model form of inter-individual and inter-occasion variability, residual error, and covariates from a common set of model options. For the simulated dataset, stepwise covariate modeling identified three of four true covariates and two spurious covariates; Lasso identified two of four true and 0 spurious covariates; and the single-objective, hybrid genetic algorithm identified three of four true covariates and one spurious covariate. For the clinical datasets, the Akaike information criterion was a median of 22.3 points lower (range of 470.5 point decrease to 0.1 point decrease) for the best single-objective hybrid genetic-algorithm candidate model versus the final manual stepwise model: the Akaike information criterion was lower by greater than 10 points for four compounds and differed by less than 10 points for three compounds. The root mean squared error and absolute mean prediction error of the best singleobjective hybrid genetic algorithm candidates were a median of 0.2 points higher (range of 38.9 point decrease to 27.3 point increase) and 0.02 points lower (range of 0.98 point decrease to 0.74 point increase), respectively, than that of the final stepwise models. In addition, the best single-objective, hybrid genetic algorithm candidate models had successful convergence and covariance steps for each compound, used the same compartment structure as the manual stepwise approach for 6 of 7 (86 %) compounds, and identified 54 % (7 of 13) of covariates included by the manual stepwise approach and 16 covariate relationships not included by manual stepwise models. The model parameter values between the final manual stepwise and best single-objective, hybrid genetic algorithm models differed by a median of 26.7 % (q1 = 4.9 % and q3 = 57.1 %). Finally, the single-objective, hybrid genetic algorithm approach was able to identify models capable of estimating absorption rate parameters for four compounds that the manual stepwise approach did not identify. The single-objective, hybrid genetic algorithm represents a general pharmacokinetic model building methodology whose ability to rapidly search the feasible solution space leads to nearly equivalent or superior model fits to pharmacokinetic data.

KW - Genetic Algorithms

KW - Model building

KW - Pharmacokinetics

UR - http://www.scopus.com/inward/record.url?scp=84865526608&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84865526608&partnerID=8YFLogxK

U2 - 10.1007/s10928-012-9258-0

DO - 10.1007/s10928-012-9258-0

M3 - Article

C2 - 22767341

AN - SCOPUS:84865526608

VL - 39

SP - 393

EP - 414

JO - Journal of Pharmacokinetics and Pharmacodynamics

JF - Journal of Pharmacokinetics and Pharmacodynamics

SN - 1567-567X

IS - 4

ER -