Bayesian versus Frequentist statistical modeling: a debate for hit selection from HTS campaigns

Area of Science:

• Pharmacology and Drug Discovery
• Statistical Methodology in Research

Background:

• The Bayesian-Frequentist debate represents fundamental differences in statistical inference and research philosophy.
• Early-stage decisions in drug discovery, including hit selection from high-throughput screening (HTS), significantly affect later development success.
• The increasing volume and complexity of data from HTS campaigns necessitate advanced statistical analysis for effective hit prioritization.

Purpose of the Study:

• To review recent statistical analysis advancements for hit selection in drug discovery.
• To explore the role and potential impact of the Bayesian-Frequentist debate on prioritizing hits from HTS campaigns.
• To assess how statistical approaches align with the growing information content in HTS data over time.

Main Methods:

• Literature review of current statistical analyses employed in drug discovery hit selection.
• Analysis of the influence of early-stage decision-making on overall HTS performance.
• Examination of the convergence and divergence of Bayesian and Frequentist approaches in the context of HTS data.

Main Results:

• Recent statistical methods are being applied to improve hit selection in drug discovery.
• Decisions made early in the drug discovery pipeline have a substantial effect on subsequent HTS performance.
• While Bayesian and Frequentist statistical approaches strive for consensus, their ability to provide identical answers may be limited as HTS data value increases.

Conclusions:

• The Bayesian-Frequentist debate is increasingly relevant to hit prioritization in HTS.
• Effective statistical analysis is crucial for optimizing drug discovery pipelines by informing early-stage decisions.
• Further research is needed to fully understand the implications of these statistical debates for successful drug development.