Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Non-linear regression analysis with errors in both variables: estimation of co-operative binding parameters.

G Valsami1, A Iliadis, P Macheras

  • 1School of Pharmacy, University of Athens, Athens, Greece.

Biopharmaceutics & Drug Disposition
|October 20, 2000
PubMed
Summary

For correlated errors in regression variables, parameter estimation methods like GMFR and ML perform poorly. For uncorrelated errors, methods show similar RMSE, but WLS and PLS offer better bias control than GMFR and ML.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Columbus' egg: Oral drugs are absorbed in finite time.

European journal of pharmaceutical sciences : official journal of the European Federation for Pharmaceutical Sciences·2022
Same author

Clamped percutaneous Poller screws to assist intramedullary nailing of proximal third tibia shaft fractures.

Annals of the Royal College of Surgeons of England·2022
Same author

Enhancement of Docetaxel Absorption Using Ritonavir in an Oral Milk-Based Formulation.

Pharmaceutical research·2021
Same author

Dose individualization of intravenous busulfan in pediatric patients undergoing bone marrow transplantation: impact and in vitro evaluation of infusion lag-time.

The Journal of pharmacy and pharmacology·2021
Same author

The effect of athletes` hyperhydration on the urinary 'steroid profile' markers in doping control analysis.

Drug testing and analysis·2018
Same author

On the pharmacokinetics of two inhaled budesonide/formoterol combinations in asthma patients using modeling approaches.

Pulmonary pharmacology & therapeutics·2017

Area of Science:

  • Biostatistics
  • Pharmacokinetics
  • Computational Chemistry

Background:

  • Parameter estimation is crucial for modeling biological systems, especially when dealing with experimental data where both variables may contain errors.
  • The Hill model is frequently used to describe co-operative drug-protein binding kinetics.
  • Choosing an appropriate parameter estimation method is vital for accurate model fitting and reliable biological interpretation.

Purpose of the Study:

  • To evaluate and compare the performance of four parameter estimation criteria: geometric mean functional relationship (GMFR), maximum likelihood (ML), perpendicular least-squares (PLS), and non-linear weighted least squares (WLS).
  • To assess these criteria when fitting the Hill model to simulated data with errors in both regression variables.
  • To determine the impact of correlated and uncorrelated errors on the accuracy and reliability of parameter estimates.

Related Experiment Videos

Main Methods:

  • Simulated data sets with known variances were generated, incorporating errors in both regression variables.
  • The Hill model for co-operative drug-protein binding was fitted to the simulated data using GMFR, ML, PLS, and WLS criteria.
  • Performance was quantified by evaluating bias, relative standard deviation (S.D.), and root-mean-squared error (RMSE) between estimated and true parameter values.

Main Results:

  • All tested criteria performed poorly with correlated errors, with GMFR and ML showing particularly weak performance.
  • For uncorrelated errors, all criteria yielded comparable root-mean-squared errors (RMSE).
  • GMFR and ML resulted in lower standard deviations (S.D.) but higher biases compared to WLS and PLS.
  • The WLS criterion showed reduced performance when equal dispersion was assumed for both observed variables.

Conclusions:

  • Parameter estimation methods exhibit varying sensitivities to error structures in regression variables.
  • For drug-protein binding models with errors in both variables, WLS and PLS may be preferred over GMFR and ML due to better bias control, especially when errors are uncorrelated.
  • Careful consideration of error characteristics is essential for selecting appropriate parameter estimation techniques in pharmacokinetic and pharmacodynamic modeling.