Applications of multivariate statistical methods to drug discovery and spectroscopy

Yegor Zyrianov, Purdue University


Modern methods of the experimental data collection, storage and processing require building an interdisciplinary connection between chemistry and multivariate statistical analysis. This work presents examples of several applications of standard statistical methods (MDS, HCA, PCA, PCR, univariate and multivariate linear regressions, distribution characterization, Least-Squares methods and others) for some chemical problems from the areas of drug design and discovery and spectroscopic data analysis. ^ First publication presents an example of treatment of spectroscopical binary data. Chapter One further introduces and develops basic approaches to handle, understand, visualize and quantify the spectral data. Chapter Two describes novel gray-system deconvolution algorithm IRONFLEA and its implementation in a computer code. Examples and issues of PCR are presented and discussed in one of the appendices. Second publication presents application of multivariate methods and quantum-mechanical calculations for the development of ROKS molecular shape descriptors. ^




Dor Ben-Amotz, Purdue University.

Subject Area

Statistics|Chemistry, Organic

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server