Beyond IC₅₀s: Towards Robust Statistical Methods for in vitro Association Studies

Area of Science:

• Pharmacology and Toxicology
• Genetics and Genomics
• Biostatistics

Background:

• Cell line cytotoxicity assays are crucial for studying differential cytotoxic responses in genetic and genomic research.
• Current studies often use curve-fitting to derive single summary measures like EC/IC50 for genetic association analyses.
• There is a need to evaluate the statistical validity of these common approaches.

Purpose of the Study:

• To critically evaluate the limitations of using EC/IC50 as a phenotype in genetic association studies.
• To highlight potential statistical shortcomings of current curve-fitting methods.
• To discuss more robust alternative analysis strategies for cytotoxicity data.

Main Methods:

• Review of statistical approaches commonly used in cell line cytotoxicity assays.
• Discussion of the assumptions and limitations inherent in EC/IC50 parameter estimation.
• Presentation of a simulation experiment to illustrate potential statistical issues.
• Exploration of alternative, more robust analytical methods.

Main Results:

• The EC/IC50 measure, while popular, is based on significant modeling assumptions and has statistical limitations when used as a phenotype.
• Simulation experiments reveal potential issues with using EC/IC50 in association studies.
• Alternative analysis methods demonstrate greater robustness.

Conclusions:

• The reliance on EC/IC50 as a sole phenotype in genetic association studies derived from cytotoxicity assays may be statistically suboptimal.
• Researchers should consider the limitations of this approach and explore more robust statistical methodologies for analyzing differential cytotoxic responses.