University of Pittsburgh researchers have developed the Pittsburgh Reproducibility Protocol (PReP), an approach which uses common statistical metrics to assess the inter- and intra-study reproducibility of microphysiological systems (MPS) studies. Designed to be included in the workflow of any MPS study, PReP can distinguish biological or clinical heterogeneity from experimental variability. Greater understanding of variability and reproducibility is vital for both basic research and precision medicine. Inclusion of PReP in workflows will improve the clinical application of research.
PReP has been developed as a workflow using statistical analysis to determine the reproducibility of preclinical research models. PReP can be applied to both single and multiple time points and assess the intra- and inter-study reproducibility in addition to distinguishing biological and clinical heterogeneity from experimental variability.
Description
MPS are used in multi-cell type, 3D non-animal models as in vitro models to study organ systems and functions, identify therapeutic strategies, and select cohorts for clinical trials. Together with organs on chips, MPS are powerful tools in preclinical and basic research enabling scientists to move away from animal testing. However, reproducibility of preclinical results is a major challenge, and it is necessary to validate any preclinical model for accuracy and reproducibility prior to being fully adopted in the drug development pipeline. PReP is a statistical tool to validate models based on experimental data and could dramatically improve preclinical reproducibility and the use of preclinical models in research.
Applications
• Preclinical studies
• Clinical trials
• Precision medicine
Advantages
Some estimates suggest irreproducibility in preclinical studies could be as high as 90%, hindering clinical applications of laboratory results. The causes of irreproducibility are varied and include poor experiment design, poorly validated methods, inappropriate statistical analysis and interpretation, and biological variation. PReP uses statistical analysis including coefficient of variation (CV), analysis of variance (ANOVA), and intraclass correlation coefficient (ICC) of large, complex datasets from preclinical studies to determine reproducibility. PReP assesses intra- and inter-study experimental variability, validating preclinical models for widespread use. PReP can also identify biological heterogeneity (i.e., inherent biological differences) aiding the development of precision therapeutics by identifying cohorts most likely to respond to treatment.
Invention Readiness
PReP has been developed and involves successive statistical analysis (CV, ANOVA, and ICC) on multiple levels of experimental data to determine reproducibility based on the results of each statistical test. PReP has been applied to establish the reproducibility of metabolic dysfunction-associated steatotic liver disease models. PReP could be applied to any preclinical model and in future could be used to validate models for approval by regulators (e.g., FDA) leading to widespread use of preclinical models in drug development.
IP Status
Copyright
