Modeling error in experimental assays using the bootstrap principle: understanding discrepancies between assays using different dispensing technologies

  • Sonya M. Hanson, Sean Ekins, John D. Chodera
  • Journal of Computer-Aided Molecular Design, December 2015, Springer Science + Business Media
  • DOI: 10.1007/s10822-015-9888-6

understanding error with different dispensing tools

What is it about?

The bootstrap principle is used to model error with different dispensing tools such as acoustic and tip-based systems. We used the data from the 2013 PLOSONE paper on the EphB4 inhibitors and showed that this data modeling approach can explain some of the error in the IC50 values obtained with both dispensing approaches.

Why is it important?

An Ipython notebook is made available so that others can adjust the parameters from their own experiments and they can see for themselves what affects assay error. The paper is important because of the detailed approach taken to getting to the bottom of assay error with different dispensing approaches.

Perspectives

Dr Sean Ekins
Collaborations in Chemistry

The contributions from this paper show how manufacturer provided precision and accuracy data can be used with bootstrapping to simulate virtual experiments. While adding the bias shifts the pIC50 data for tips and acoustic approaches closer to equivalence it leaves the door open to what else might be contributing.

Read Publication

http://dx.doi.org/10.1007/s10822-015-9888-6

The following have contributed to this page: Dr Sean Ekins