Objectives
Creation of a common standard for NMR-spectra and related experimental data
Creation of procedures and algorithms for the automatic inclusion of NMR spectra P2P network (.
Development of algorithms for automatic NMR-spectra interpretation.
Feasibility study of the reliably of the automatic assignment procedures.
Preparation of manuals and tutorials for NMR tools and databases.
Current status
Nuclear Magnetic Resonance (NMR) measurements enable a complete assignment of all structural details of complex carbohydrates. The lack of appropriate reference data complied in well-structured data collections as well as the deficiency of well-evaluated algorithms, which have been tested for various classes of carbohydrates, are the main obstacles for a fully automatic interpretation of NMR spectra. It is the aim of DS4-T1 and DS4-T2 to overcome this unfortunate situation by providing conventions of a uniform representation of NMR data. Equally important are agreements on quality measures of NMR spectra, which have to be fulfilled before entering the database. DS4-T3 tries to overcome this disadvantageous situation.
Currently, only a few isolated approaches exist to collect NMR spectra and to provide programs for automatic spectra interpretation. A database of about 2.000 spectra of N- and O-glycans (SugaBase) was collected at the Bijvoet Center, Utrecht (The Netherlands). However, this database is no longer updated. The University of Stockholm provides a computer program, for sequence determination of bacterial oligosaccharides and regular polysaccharides. It is the aim of DS4-T4 to expand the rule based approach developed at SU to additional classes of biologically important glycans. A Russian group is currently collecting NMR spectra of bacterial polysaccharides. However, these spectra are currently not available on the Internet. The Glycosciences.de portal maintained by DKFZ, Heidelberg provides an estimation of NMR chemical shifts for complex carbohydrates using an encoding of chemical environments. Artificial Neural Networks (ANNs) have long been discussed to be used for automatic interpretation of NMR-spectra, since they are potent tools for classification. ANNs allow extracting the correlation of structural and spectroscopic data contained in carefully assigned NMR spectra. The University of Basel has recently made available a preliminary version of an ANN approach for the analysis of 13C-NMR data of mammalian oligosaccharides and will continue to develop their approach as part of DS4-T4.
All described projects suffer from the lack of available, carefully assigned NMR spectra, the difficulties to exchange the data and the unavailability of appropriate algorithms for automatic spectrum interpretation. The integration of definitions and agreements (DS4-T1, DS4-T2 and DS4-T3) and new tools to interpret NMR spectra (DS4-T4) into the P2P network (DS4-T6) and its evaluation will guarantee that all procedures work smoothly so that a fully functional prototype as part of the P2P network will be available.
EuroCarbDB is a Research Infrastructure Design Study
Funded by the 6th Research Framework Program of the European Union
(Contract: RIDS Contract number 011952)
