Recently I published a paper on the estimation of IC50 values. The central idea is that it is better to estimate the sensitivity of all items (here, cell lines / compounds) simultaneously compared to one-by-one. This allows us to borrow strength across all the observations and thereby improve stability and accuracy of the estimates. This page intends to give a high-level overview of the proposed method, the peer reviewed publication can be found here (PMID 27180993, Pharmacogenomics, May 2016, Vol. 17, No. 7, Pages 691-700). The associated script/computer code can be found here.