ANYL: New Directions in Chemometrics: Making Sense of Big & Small Chemical Data Sets

This session was originally proposed as a CINF session dealing with computational methods for spectral analysis and spectral databases. Unfortunately, there were insufficient submissions for CINF to have a standalone session so we cosponsored and merged our submissions with those of the Analytical Chemistry Division. Although the session was scheduled for Thursday, including Thursday afternoon, the symposium was very well attended (thanks to the University of Delaware graduate students).

The use of the KNIME pipelining analytics platform for processing metabolomics MS/MS data performing noise reduction, signal normalization, peak picking, retention time correction, isotope merging, and annotation was presented. Some of the symposium was concerned with developing a modeling strategy that is robust to spectral interferents based on adaptive regression and multivariate calibration. Dr. J. Johnson from the Hilmar cheese company in California discussed the chemometric models used for real-time spectral monitoring of cheese processing for the FDA. Dr. Curtis Mowry of Sandia National Labs discussed principal component analysis of X-ray fluorescence (XRF) iron andc hromium peaks to obtain a signature to identify adulterated materials. Dr. Douglas Van der Gried discussed obtaining binding constants from spectrophotometric data and methods for modeling spectrophotometric titration data. Dr. Joseph Smith discussed Raman microspectroscopic data mapping. Dr. Barry Lavine presented principal component analysis methods of MALDI/MS-MS profiling for altered glycans as a cancer diagnostic. Dr. Kevin Moore of USP Pharmacopeia discussed the analysis of FTIR data of excipients (the “inert” drug ingredients). Dr. Greg Banik discussed Bio-Rad’s spectral optimization and correction methods and spectral identification databases. Dr. Tony Williams presented the new EPA ToxCast Interactive Chemical Safety for Sustainability (iCSS) dashboard available for public use and data access and analysis. Dr. Pillhum Son of CAS closed out the session presenting spectral data available through the CAS REGISTRY databases.

All in all, many useful methods and algorithms for spectral analysis and spectral identification through database comparison were presented, as well as the application of spectral data for identification of sample quality, defects, and alteration.

Rachelle Bienstock