Overview

The project as a whole seems to be progressing well, although the NMR part of the project lags a little behind the HPLC and MS aspects. Part of the discussions were naturally concerned with the premature passing of their principle coordinator Willi von der Lieth.

The extent to which CCPN was accepted was impressive. Everyone was very happy with the ChemComps and it was generally accepted that all NMR data would be under the CCPN model. It was timely that I presented Analysis at this meeting as the assignment process needs to be addressed..

Scientifically the emphasis seemed to be moving toward the analysis of mixtures of carbohydrates and situations where the covalent structure is unknown. NMR takes a key part in this and people want to us it as a tool to determine which types of residues are in their sample, something that us protein people should be mindful of.

Presentations

Matt Harrison presented the core part of the database. He demonstrated a working example filled in from existing carbohydrate data sources and showed the interface to add new data, make queries etc. The system looked impressive with neat covalent structure representation for the carbohydrate residues and seemed powerful yet simple for the everyday things.

Thomas Lutteke presented the Monocaccharide DB, which seemed to be nearly ready. As with some of the other software, his interface gets the user to think of which residues they want in terms of core ring + stereochemistry + conjugation +  reducing end. They were fully aware of outputting everything as ChemComps, the core part of which should serve as the basis of carbohydrate templates that we can distribute with CCPN.

Magnus Lundberg presented CASPER, a program which can both predict chemical shifts given an input covalent structure and predict covalent structure given chemical shifts. Magnus gave a demonstration of a web interface that and showed some impressively accurate results. The CASPER program, via the C API, can both accept data from a CCPN project (uploaded as .tar.gz) and send it back to a CCPN project. Accordingly Magnus nicely demonstrated how the CASPER output can be directly visualised in Analysis: A nicely convincing example of why carbohydrate NMR people should use Analysis.

Ana Arda-Freire described the NMR workflow, partly for the benefit of the non-NMR people. The carbohydrate ChemComps and Analysis were an accepted part of this ans it seems that she has been successfully using Analysis for carbohydrate assignment for a while now.

Bas Leeflang presented the ProSpectND processing package. The program seemed neat and user friendly. We may want to look at the interface when it comes to CcpNmr Processing.

I presented Analysis and might have gained some new users in the labs of Javier Canada and Marco Guerrini respectively. Javier was having a hard time with Cara and seemed converted, plus we might get invited to Madrid.

Specific Development Points

• The CCPN Java SQL API is eagerly awaited as they are relying on this bing ready to help the NMR part of the project catch up with the MS/HPLC part.
• There was a request for unknown and partially unknown carbohydrate ChemComps, e.g. "hexose". I don't know if there are already plans for this.
  
• There is a Python Glyco-CT reader which Matt Harrison has; I have said that I will see if this can easily be used to build molecules in Analysis, which have come from their swish graphical carbohydrate builder.
  
• Couplings are very important and frequently present in the carbohydrate spectra, it will be important to make sure that the next Analysis version handles this.
• There was a request for pseudo-2D handling (which might mostly work already) to measure build-up curves etc., e.g. DOSY. I shall ask for some example spectra for this.
  
• Bitmap file export of spectra, which could be automated, was requested.
• We need to improve the labelling of 1D spectra for multiplets.
• Resolving ambiguous MolResidue types in Analysis would be useful for the covalent structure elucidation process.
• CASPER could in the future be linked directly to Analysis, but we need to think of the best way to work between the Python and C APIs.
• Fanny Bonachera from a group in Lille seems keen to use the Python modules from Analysis (both analysis.*Basic and memops.gui) for a project involving the comparison of spectra: I'm not sure what the details of this are but it may well be something to do with resolving mixtures. In any case I am in contact and can point them in the right direction when I know more. Given that they are in Lille they can easily visit.