%0 Journal Article
%T Distributing Correlation Coefficients of Linear Structure-Activity/Property Models
%A Lorentz J£¿NTSCHI
%A Sorana D. BOLBOACA
%J Leonardo Journal of Sciences
%D 2011
%I AcademicDirect
%X Quantitative structure-activity/property relationships are mathematical relationships linking chemical structure and activity/property in a quantitative manner. These in silico approaches are frequently used to reduce animal testing and risk-assessment, as well as to increase time- and cost-effectiveness in characterization and identification of active compounds. The aim of our study was to investigate the pattern of correlation coefficients distribution associated to simple linear relationships linking the compounds structure with their activities. A set of the most common ordnance compounds found at naval facilities with a limited data set with a range of toxicities on aquatic ecosystem and a set of seven properties was studied. Statistically significant models were selected and investigated. The probability density function of the correlation coefficients was investigated using a series of possible continuous distribution laws. Almost 48% of the correlation coefficients proved fit Beta distribution, 40% fit Generalized Pareto distribution, and 12% fit Pert distribution.
%K Probability density function
%K Determination coefficient
%K Quantitative Structure-Activity/Property Relationship (QSAR/QSPR)
%K Molecular Descriptor Family on Vertices (MDFV)
%U http://ljs.academicdirect.org/A20/027_048.pdf