A comparison and chemometric analysis of several molecular mechanics forcefields and parameter sets applied to carbohydrates

Citation
S. Perez et al., A comparison and chemometric analysis of several molecular mechanics forcefields and parameter sets applied to carbohydrates, CARBOHY RES, 314(3-4), 1998, pp. 141-155
Citations number
33
Categorie Soggetti
Agricultural Chemistry","Chemistry & Analysis","Organic Chemistry/Polymer Science
Journal title
CARBOHYDRATE RESEARCH
ISSN journal
00086215 → ACNP
Volume
314
Issue
3-4
Year of publication
1998
Pages
141 - 155
Database
ISI
SICI code
0008-6215(199812)314:3-4<141:ACACAO>2.0.ZU;2-#
Abstract
Carbohydrates are thought to be especially difficult to model because of th eir highly polar functionality, their flexibility, and their differences in electronic arrangements that occur during conformational and configuration al changes, such as the anomeric, exo-anomeric and gauche effects. These is sues have been addressed in recent years, yielding several contributions to set up some relevant parameterizations that would account for these specif ic features of carbohydrates. Within the framework of a workshop involving the participation of 11 research groups active in the field, several common ly used molecular mechanics force fields and special carbohydrate parameter sets have been considered. The application of 20 force fields and/or sets of parameters to a series of seven test cases provided a fairly general pic ture of the potentiality of these parameter sets for giving a consistent im age of structure and energy of carbohydrate molecules. The results derived from a chemometric analysis (principal component analysis, PCA) give a glob al view of the performances of the force fields and parameter sets for carb ohydrates. The present analysis (i) provides an identification of the param eter sets which differ from the bulk, (ii) helps to establish the relations hip that exists between the different parameter sets, (iii) provides indica tions for selecting different parameter sets to explore the force held depe ndency (or the lack of thereof) of a given molecular modeling study. Throug h the PCA, we have created a force field landscape on which the different f orce fields are related to each other on a relative scale. New carbohydrate force fields can easily be inserted inter this landscape (PCA model) and r elated to the performance of existing force fields. (C) 1998 Elsevier Scien ce Ltd. All rights reserved.