Self-consistency analysis of physical property and molecular descriptor databases using a variety of prediction techniques - Id# 255431

Mordechai Shacham, Michael Elly, Inga Paster, Neima Brauner

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The dominant descriptor version of the targeted QSPR method (TQSPR1, Shacham and Brauner, Chem. Eng. Sci., 66, 2606, 2011) is used to predict 16 constant properties for close to 1800 compounds in order to assess the range of applicability of the prediction technique and to analyze the consistency of a property and molecular descriptor database. It is demonstrated that the TQSPR1 method can model the properties within the data uncertainty level for most groups of compounds included in the database. Common causes of poor prediction accuracy were identified as: 1. inconsistencies in the 3D molecular structure of the compounds; 2. use of molecular descriptors outside their range of applicability; 3. mixing compounds at different phases for properties defined at a standard state and 4. use of descriptors whose asymptotic behavior do not match that of the modeled property for long range extrapolation to high carbon number compounds. The results of this study enable improving the consistency of the physical property and descriptor databases for increasing the robustness of QSPRs that can be derived for a variety of properties and compounds.

Original languageEnglish
Title of host publicationAIChE 2012 - 2012 AIChE Annual Meeting, Conference Proceedings
StatePublished - 1 Dec 2012
Event2012 AIChE Annual Meeting, AIChE 2012 - Pittsburgh, PA, United States
Duration: 28 Oct 20122 Nov 2012

Publication series

NameAIChE Annual Meeting, Conference Proceedings

Conference

Conference2012 AIChE Annual Meeting, AIChE 2012
Country/TerritoryUnited States
CityPittsburgh, PA
Period28/10/122/11/12

Fingerprint

Dive into the research topics of 'Self-consistency analysis of physical property and molecular descriptor databases using a variety of prediction techniques - Id# 255431'. Together they form a unique fingerprint.

Cite this