Experimental design and measurement uncertainty in ligand binding studies by affinity capillary electrophoresis

Matthias Stein, Rob Haselberg, Mona Mozafari-Torshizi, Hermann Wätzig

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

In all life sciences ligand binding assays (LBAs) play a crucial role. Unfortunately these are very error prone. One part of this uncertainty results from the unavoidable random measurement uncertainty, another part can be attributed to the experimental design. To investigate the latter, uncertainty propagation was evaluated as a function of the given experimental design. A design space including the normalized maximum response range (nMRR), the data point position (DPP), the data point range (DPR) and the number of data points (NoDP) was defined. Based on ten measured ms ACE source data sets 20 specific parameter sets were selected by Design of Experiments. Monte Carlo simulations using 100 000 repeats for every parameter set were employed. The resulting measurement uncertainty propagation factors (measurement uncertainty multiplier: MUM) were used to describe the whole design space by polynomial regression. The resulting 5-dimensional response surface was investigated to evaluate the design parameter's influence and to find the minimal uncertainty propagation. It could be shown, that the nMRR is of highest importance, followed by DPP and DPR. Interestingly, the NoDP is less relevant. However, the interactions of the four parameters need to be carefully considered during design optimization. Using at least five data points which cover over 40% of the upper part of the binding hyperbola (DPP > 0.57) the MUM will be minimized (MUM approximately 1.5) when the nMRR is appropriate. It is possible to reduce the measurement uncertainty propagation more than one order of magnitude.

LanguageEnglish
Pages1-14
JournalElectrophoresis
DOIs
Publication statusPublished - 14 Jan 2019

Fingerprint

Capillary electrophoresis
Capillary Electrophoresis
Design of experiments
Uncertainty
Research Design
Ligands
Biological Science Disciplines
Assays
Polynomials

Keywords

  • Affinity capillary electrophoresis
  • Drug binding studies
  • Experimental design
  • Ligand binding assay
  • Measurement uncertainty

Cite this

@article{663ad1c394ca417eb8edb905d5e9cec3,
title = "Experimental design and measurement uncertainty in ligand binding studies by affinity capillary electrophoresis",
abstract = "In all life sciences ligand binding assays (LBAs) play a crucial role. Unfortunately these are very error prone. One part of this uncertainty results from the unavoidable random measurement uncertainty, another part can be attributed to the experimental design. To investigate the latter, uncertainty propagation was evaluated as a function of the given experimental design. A design space including the normalized maximum response range (nMRR), the data point position (DPP), the data point range (DPR) and the number of data points (NoDP) was defined. Based on ten measured ms ACE source data sets 20 specific parameter sets were selected by Design of Experiments. Monte Carlo simulations using 100 000 repeats for every parameter set were employed. The resulting measurement uncertainty propagation factors (measurement uncertainty multiplier: MUM) were used to describe the whole design space by polynomial regression. The resulting 5-dimensional response surface was investigated to evaluate the design parameter's influence and to find the minimal uncertainty propagation. It could be shown, that the nMRR is of highest importance, followed by DPP and DPR. Interestingly, the NoDP is less relevant. However, the interactions of the four parameters need to be carefully considered during design optimization. Using at least five data points which cover over 40{\%} of the upper part of the binding hyperbola (DPP > 0.57) the MUM will be minimized (MUM approximately 1.5) when the nMRR is appropriate. It is possible to reduce the measurement uncertainty propagation more than one order of magnitude.",
keywords = "Affinity capillary electrophoresis, Drug binding studies, Experimental design, Ligand binding assay, Measurement uncertainty",
author = "Matthias Stein and Rob Haselberg and Mona Mozafari-Torshizi and Hermann W{\"a}tzig",
year = "2019",
month = "1",
day = "14",
doi = "10.1002/elps.201800450",
language = "English",
pages = "1--14",
journal = "Electrophoresis",
issn = "0173-0835",
publisher = "Wiley-VCH Verlag",

}

Experimental design and measurement uncertainty in ligand binding studies by affinity capillary electrophoresis. / Stein, Matthias; Haselberg, Rob; Mozafari-Torshizi, Mona; Wätzig, Hermann.

In: Electrophoresis, 14.01.2019, p. 1-14.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - Experimental design and measurement uncertainty in ligand binding studies by affinity capillary electrophoresis

AU - Stein, Matthias

AU - Haselberg, Rob

AU - Mozafari-Torshizi, Mona

AU - Wätzig, Hermann

PY - 2019/1/14

Y1 - 2019/1/14

N2 - In all life sciences ligand binding assays (LBAs) play a crucial role. Unfortunately these are very error prone. One part of this uncertainty results from the unavoidable random measurement uncertainty, another part can be attributed to the experimental design. To investigate the latter, uncertainty propagation was evaluated as a function of the given experimental design. A design space including the normalized maximum response range (nMRR), the data point position (DPP), the data point range (DPR) and the number of data points (NoDP) was defined. Based on ten measured ms ACE source data sets 20 specific parameter sets were selected by Design of Experiments. Monte Carlo simulations using 100 000 repeats for every parameter set were employed. The resulting measurement uncertainty propagation factors (measurement uncertainty multiplier: MUM) were used to describe the whole design space by polynomial regression. The resulting 5-dimensional response surface was investigated to evaluate the design parameter's influence and to find the minimal uncertainty propagation. It could be shown, that the nMRR is of highest importance, followed by DPP and DPR. Interestingly, the NoDP is less relevant. However, the interactions of the four parameters need to be carefully considered during design optimization. Using at least five data points which cover over 40% of the upper part of the binding hyperbola (DPP > 0.57) the MUM will be minimized (MUM approximately 1.5) when the nMRR is appropriate. It is possible to reduce the measurement uncertainty propagation more than one order of magnitude.

AB - In all life sciences ligand binding assays (LBAs) play a crucial role. Unfortunately these are very error prone. One part of this uncertainty results from the unavoidable random measurement uncertainty, another part can be attributed to the experimental design. To investigate the latter, uncertainty propagation was evaluated as a function of the given experimental design. A design space including the normalized maximum response range (nMRR), the data point position (DPP), the data point range (DPR) and the number of data points (NoDP) was defined. Based on ten measured ms ACE source data sets 20 specific parameter sets were selected by Design of Experiments. Monte Carlo simulations using 100 000 repeats for every parameter set were employed. The resulting measurement uncertainty propagation factors (measurement uncertainty multiplier: MUM) were used to describe the whole design space by polynomial regression. The resulting 5-dimensional response surface was investigated to evaluate the design parameter's influence and to find the minimal uncertainty propagation. It could be shown, that the nMRR is of highest importance, followed by DPP and DPR. Interestingly, the NoDP is less relevant. However, the interactions of the four parameters need to be carefully considered during design optimization. Using at least five data points which cover over 40% of the upper part of the binding hyperbola (DPP > 0.57) the MUM will be minimized (MUM approximately 1.5) when the nMRR is appropriate. It is possible to reduce the measurement uncertainty propagation more than one order of magnitude.

KW - Affinity capillary electrophoresis

KW - Drug binding studies

KW - Experimental design

KW - Ligand binding assay

KW - Measurement uncertainty

UR - http://www.scopus.com/inward/record.url?scp=85061429021&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85061429021&partnerID=8YFLogxK

U2 - 10.1002/elps.201800450

DO - 10.1002/elps.201800450

M3 - Article

SP - 1

EP - 14

JO - Electrophoresis

T2 - Electrophoresis

JF - Electrophoresis

SN - 0173-0835

ER -