The Story

This research indicates that the responses of the components in many complex natural mixtures have a log-normal distribution. This suggests that the distributions of both the concentrations and the response factors follow a mathematical function. This function of concentration distributions had been assumed to be exponential but without confirming empirical evidence.

Luc Nagels published the paper incorrectly cited as confirmation of the exponential distribution model. I contacted him for his original data on plant lipid mixtures and we ended up working together to reexamine his data to find a distribution model that would fit. The result was a strong fit to a distribution called log-normal. This is a normal distribution of the logs of the responses.

Knowing the distribution of responses aids in determining how many components a comprehensive (untargeted) analysis may be missing. The information available from determining this distribution is shown in the illustration for this category. I undertook this work to demonstrate that the instrumental factor limiting the number of components detected is dynamic range.

A log-normal distribution of responses requires that there be patterns to the distributions of the concentrations and the response factors. If the concentrations have a log-normal distribution, there are several possible distributions of the response factor as developed in the paper with Alex.

There is more work to be done in showing the advantages of having this information about the complex mixtures one is studying.

2011 Undetected Components in Natural Mixtures
2011 Undetected Components, supporting document
2015 Potential Response Factor Distributions

Just out, but not yet in print, is my paper intitled, “Using the Response Distribution to Compare and optimize Untargeted Analysis Techniques.”

If you are looking for the spreadsheet to perform the log-normal characterization of response data, double-click here.

Papers on out study of the composition of complex natural mixtures. Collaborators in this work were Luc Nagels, University of Antwerp, and Alexander Gundlach-Graham, Iowa State University.