Analysis of small-angle scattering data using model fitting and Bayesian regularization
What is it about?
The paper shows how prior knowledge can be directly included in the fitting process of small-angle scattering data. Prior knowledge is all the knowledge about the system that is available before (prior to) the experiment. That can include independent measurements, e.g. of the protein concentration, or of the number of phospholipids in the sample. It can also include knowledge about the some of the molecules in the sample. That could be the expected width of the alpha-helices, or the headgroup area of the lipids. It could also be the electron density of the different components. All this information should be taken into account when analysing the data. We show in the paper how a Bayesian methods provide the perfect tool for inclusion of the prior knowledge, and gives two experimental examples. The paper also introduces a measure for the information in data, i.e. how many effective parameters can be determined from the data, and discuss this in relation to other measures for the information content in small-angle scattering data.