The development of high-throughput assays is typically the bottleneck in the high-through screening process.  Assays for high-throughput screening must have sufficient robustness that they do not generate a large number of false-positive results. A typical lab assay may be considered adequate even if it shows significance with an alpha level of 0.05. However, applied to a primary screen of 100000 compounds, this would generate an unacceptable 5000 false positives.

A commonly applied rule in HTS is that the assay should routinely exhibit a z’ factor of >0.5 which amounts to >12 standard deviations difference between relevant positive and negative controls (1). Although this requirement may be loosened in some cases, it is also important that this robustness is maintained reproducibly on multiple assay plates and each data the assay is used (2).

The National Centre for Advancement of Translational Studies (NCATS) have developed a guide for evaluating whether an assay passes the criteria required for a high-throughput screen (2).

The screening unit can offer advice on developing assay robustness and evaluating it according to this NCATS standard (2). Below are the results of one of the screening unit’s assays, run on a 384 well plate with positive, negative and “half-way point” inhibitor samples in 3 different plate arrangments and on 3 separate days. Some variation can be seen, especially from day to day, and even the fractional effect of the nominal half-way point varies. The differences are nevertheless more than sufficient to identify hits in an efficient way with minimal false-positives and passes all criteria specified by the test (2).

Below is the result obtained at our unit (by Arkadiusz Popinigis) using the screen above when applied to ~750 compounds that were cherry picked from in silico screens performed by Tuomo Laitinen and Katryna Cisek.

Below is a side-by-side comparison of the assay used with other commonly used protein interaction methods, fluorescent polarisation and microscale thermophoresis, applied to the same interaction (by Li-Li Li and Jesse Mattson). As indicated the quantity of protein required per assay is extremely low, minimising the effort needed to run a screen, while the z’ factor is notably higher than the other methods.

1. Zhang JH, Chung TD, Oldenburg KR. A Simple Statistical Parameter for Use in Evaluation and Validation of High Throughput Screening Assays. J Biomol Screen. 1999;4(2):67-73. PubMed PMID: 10838414.
2. Iversen PW, Beck B, Chen YF, et al. HTS Assay Validation. 2012 May 1 [Updated 2012 Oct 1]. In: Sittampalam GS, Grossman A, Brimacombe K, et al., editors. Assay Guidance Manual [Internet]. Bethesda (MD): Eli Lilly & Company and the National Center for Advancing Translational Sciences; 2004-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK83783