Corinna Maier1,2, Carolin Loos1,2, Jan Hasenauer1,2. 1. Helmholtz Zentrum München - German Research Center for Environmental Health, Institute of Computational Biology, Neuherberg 85764, Germany. 2. Technische Universität München, Center for Mathematics, Chair of Mathematical Modeling of Biological Systems, Garching 85748, Germany.
Abstract
Motivation: Dynamics of cellular processes are often studied using mechanistic mathematical models. These models possess unknown parameters which are generally estimated from experimental data assuming normally distributed measurement noise. Outlier corruption of datasets often cannot be avoided. These outliers may distort the parameter estimates, resulting in incorrect model predictions. Robust parameter estimation methods are required which provide reliable parameter estimates in the presence of outliers. Results: In this manuscript, we propose and evaluate methods for estimating the parameters of ordinary differential equation models from outlier-corrupted data. As alternatives to the normal distribution as noise distribution, we consider the Laplace, the Huber, the Cauchy and the Student's t distribution. We assess accuracy, robustness and computational efficiency of estimators using these different distribution assumptions. To this end, we consider artificial data of a conversion process, as well as published experimental data for Epo-induced JAK/STAT signaling. We study how well the methods can compensate and discover artificially introduced outliers. Our evaluation reveals that using alternative distributions improves the robustness of parameter estimates. Availability and Implementation: The MATLAB implementation of the likelihood functions using the distribution assumptions is available at Bioinformatics online. Contact: jan.hasenauer@helmholtz-muenchen.de. Supplementary information: Supplementary material are available at Bioinformatics online.
Motivation: Dynamics of cellular processes are often studied using mechanistic mathematical models. These models possess unknown parameters which are generally estimated from experimental data assuming normally distributed measurement noise. Outlier corruption of datasets often cannot be avoided. These outliers may distort the parameter estimates, resulting in incorrect model predictions. Robust parameter estimation methods are required which provide reliable parameter estimates in the presence of outliers. Results: In this manuscript, we propose and evaluate methods for estimating the parameters of ordinary differential equation models from outlier-corrupted data. As alternatives to the normal distribution as noise distribution, we consider the Laplace, the Huber, the Cauchy and the Student's t distribution. We assess accuracy, robustness and computational efficiency of estimators using these different distribution assumptions. To this end, we consider artificial data of a conversion process, as well as published experimental data for Epo-induced JAK/STAT signaling. We study how well the methods can compensate and discover artificially introduced outliers. Our evaluation reveals that using alternative distributions improves the robustness of parameter estimates. Availability and Implementation: The MATLAB implementation of the likelihood functions using the distribution assumptions is available at Bioinformatics online. Contact: jan.hasenauer@helmholtz-muenchen.de. Supplementary information: Supplementary material are available at Bioinformatics online.
Authors: Laura Marie Helleckes; Michael Osthege; Wolfgang Wiechert; Eric von Lieres; Marco Oldiges Journal: PLoS Comput Biol Date: 2022-03-07 Impact factor: 4.475
Authors: Tomer Zohar; Carolin Loos; Stephanie Fischinger; Caroline Atyeo; Chuangqi Wang; Matthew D Slein; John Burke; Jingyou Yu; Jared Feldman; Blake Marie Hauser; Tim Caradonna; Aaron G Schmidt; Yongfei Cai; Hendrik Streeck; Edward T Ryan; Dan H Barouch; Richelle C Charles; Douglas A Lauffenburger; Galit Alter Journal: Cell Date: 2020-11-03 Impact factor: 41.582