Related Experiment Video
Updated: Dec 5, 2025
11:13
Identification of Kinase-substrate Pairs Using High Throughput Screening
Published on: August 29, 2015
8.5K
Statistical models for identifying frequent hitters in high throughput screening
Samuel Goodwin1, Golnaz Shahtahmassebi1, Quentin S Hanley2
1School of Science and Technology, Nottingham Trent University, Clifton Lane, Nottingham, NG11 8NS, UK.
Scientific Reports
|October 15, 2020
Summary
The binomial survivor function model for frequent hitters in high throughput screening (HTS) identified too many inactive compounds. Alternative models, like the gamma distribution, provided a more accurate assessment of compound activity in drug discovery.
Area of Science:
- Drug discovery and development
- Computational chemistry
- Pharmacology
Background:
- High throughput screening (HTS) is crucial for identifying active compounds from large libraries.
- Existing models, like the binomial survivor function (BSF), may inaccurately classify compound activity.
- Understanding compound behavior in HTS is essential for efficient drug discovery.
Purpose of the Study:
- To evaluate the effectiveness of the binomial survivor function (BSF) model for identifying frequent hitters in HTS.
- To investigate alternative statistical models for characterizing compound behavior in HTS data.
- To analyze the implications of disproportionate compound retesting on drug discovery strategies.
Main Methods:
- Analysis of 872 publicly available HTS datasets.
- Assessment of the binomial survivor function (BSF) model.
- Investigation of generalized logistic, gamma, and negative binomial distributions as alternative models.
- Evaluation of compound testing frequency and its impact on results.
Main Results:
- The BSF model identified a large proportion of 'infrequent hitters,' leading to its rejection for frequent hitter identification.
- The gamma model reduced the proportion of both frequent and infrequent hitters compared to the BSF.
- Disproportionate retesting (≥300 times for 17.6% of compounds) dominated the datasets, suggesting compound repurposing over novel drug discovery.
- Individual compounds were poorly characterized due to limited testing, while assays were well-characterized by numerous compounds.
Conclusions:
- The BSF model is inadequate for identifying frequent hitters in HTS.
- Alternative models, such as the gamma distribution, offer improved characterization of compound behavior.
- Extensive retesting in HTS datasets may represent large-scale compound repurposing rather than early-stage drug discovery.
- The current HTS approach poorly characterizes individual compounds, necessitating a re-evaluation of drug discovery strategies.
