Quantifying assays: A Modeling tale of variability in cancer therapeutics assessed on cancer cells

TL;DR

This paper introduces a mathematical modeling approach to quantify and reduce variability in cancer cell assay data, specifically analyzing inhibition mechanisms in melanoma cells to improve the reliability of experimental inferences.

Contribution

It proposes a theoretical model of cell viability as a function of time and parameters, enhancing data interpretation and reducing variability effects in cancer therapeutics assays.

Findings

Model reduces sensitivity to data variability

Parameters directly interpret biological processes

Improves understanding of inhibition dynamics

Abstract

Inhibiting a signalling pathway concerns controlling the cellular processes of a cancer cell's viability, cell division, and death. Assay protocols created to see if the molecular structures of the drugs being tested have the desired inhibition qualities often show great variability across experiments, and it is imperative to diminish the effects of such variability while inferences are drawn. In this paper we propose the study of experimental data through the lenses of a mathematical model depicting the inhibition mechanism and the activation-inhibition dynamics. The method is exemplified through assay data obtained from the study of inhibition of the CXCL12/CXCR4 activation axis for the melanoma cells. To mitigate the effects of the variability of the data on the cell viability measurement, the cell viability is theoretically constructed as a function of time depending on several parameters. The values of these parameters are estimated by using the experimental data. Deriving approximation for the cell viability in a theoretically pre-determined form has the advantages of (i) being less sensitive to data variability (ii) the estimated values of the parameters are interpreted directly in the biological processes, (iii) the amount of variability explained via the approximation validates the quality of the model, (iv) with the data integrated into the model one can derive a more complete view over the whole process. These advantages are demonstrated in the step-by-step implementation of the outlined approach.