Drug Discovery 2022: driving the next life science revolution

AI Driven Automation of Model Selection and Data Quality Control in SPR Production Screens


P Sharma1; J Florez2; A Singh1; D Siegismund2; M Wieser2; M Pfreundschuh2; Q Chen1; C Ch'ng1; D Hammond3; S Heyse2S Steigele2
1 Amgen, UK;  2 Genedata AG, Switzerland;  3 Genedata UK, UK


The advent of automated laboratory and data analytics has increased the success of drug discovery programs, by optimizing experimental workflows to better control result quality with less investment of resources and time. However, biophysical assays such as Surface Plasmon Resonance (SPR) still binding require tedious manual review for quality control, result review and decision-making, due to the complex nature of sensorgrams and possible outcomes. Genedata, together with Amgen, developed an AI-driven data analysis workflow for automation of complex biophysical analysis. In this workflow, sensorgrams are classified automatically into four categories before the appropriate 1:1 binding models (kinetic or steady-state) are applied. This workflow ensures that binding affinity and kinetic parameters are reproducibly and precisely determined for the multitude of outcomes typically observed in a compound screen, reducing the need for expert review to just a very few corner cases. We explain the basic elements of this new AI-driven solution for automating SPR data analysis, illustrate its use on two Amgen production screens, and conclude with a discussion on its potential impact for future drug screening programs.

Hosted By


The European Laboratory Research & Innovation Group Our Vision : To provide outstanding, leading edge knowledge to the life sciences community on an open access basis