An improved method for the analysis of data from radiation-inactivation studies.

This paper addresses statistical issues in the estimation of protein molecular weight using radiation-inactivation assays. In particular it considers experiments in which a number of internal standards are used to supplement or replace accurate measurement of the applied doses of radiation. A mathematical model is proposed which allows the use of the standard technique of maximum-likelihood estimation to estimate the unknown molecular weight without knowledge of the applied doses and, in contrast to previous methods, allows the construction of confidence intervals.