I’ve been in a number of conversations of late about how Mass Spectrometrists might use ChemSpider and get value from our efforts. I recently gave a short Powerpoint presentation to a group about what ChemSpider is and the types of queries that ChemSpider users can conduct today. I’ve posted the presentation to Slideshare as usual so people can access it there if they are interested.

I’ve started wrapping my head around how we could provide more value to some of our users in regards to MS, HPLC and NMR. One of the things we could do is to use our known text mining skills to look for NMR or MS (LCMS) articles based on the use of the terms in the title or abstract and then using those terms as tags against chemical structures in the abstract/title. So, from titles such as “High-Performance Liquid Chromatographic Method for Determination of Phenytoin in Rabbits Receiving Sildenafil” from our collaborator Libertas Academica we would extract HPLC and Phenytoin and connect the article to the structure as we have done here. In this way the article would be searchable by structure and associated analytical technique and we could even look at extracting the detailed experimental approach from Open Access articles. More work but feasible. Any comments???

Stumble it!

4 Responses to “How Mass Spectrometrists Might Use ChemSpider”

  1. Chris Singleton says:

    I love the presentation, it really shows the capability of Chemspider when it comes to analytical searching. Is it possible to include/exclude moeities as well as elements? That was I can select for a particular group for a certain MS mode (carboxylic acid for negative mode for Atmospheric Pressure Ionization, for example)?
    Only thing about extracting the experimental approach, is the text mining pretty reliable? Perhaps it would be easier to include the experimental section of the paper in its entirety, rather than pull out the relevant parts for that particular experiment. I know it’s not as elegant, but I’m just wondering about how much effort we’ll need to do in manual curation (if any) if we try to extract only the relevant parts.

  2. Antony Williams says:

    Thanks for the acknowledgment re. the presentation. It’s certainly only showing some of the capabilities. It is NOT yet possible to exclude moeities as well as elements but we are considering doing exactly that. It would be good to align your expectations with our overall development plan so if you could outline your expectations for us that would be great.
    We “could” pull out the entire experimental part but I am concerned about overwhelming records. We have expanded the schema to include documents and this will make the connections much better if we do this moving forward.

  3. Ilmari Krebs says:

    A similiar question was on the tip of my tongue last week already. I was wondering if Chemspider offers searching by screen numbers as CAS uses for restricting its substructure searches. From MS point of view one would derive certain functional groups from the spectroscopic data. Searching for a certain formula and filtering the data for existance of these observed functional groups would narrow down the possiblities, hopefully.

  4. Chris Singleton says:

    I suppose my expectations would be that we can include moeties in a search as we do other physical properties. Such as MW 370, logP from 1-3, with an -NH2 group, like Ilmari suggests. If we could put it into general search terms I think it would be useful for not only for mass spectrocopists, but also the Chemspider community as a whole. Perhaps we could just add qualifiers to the existing substructure search? For example, when you search for a substructure you can have it filter on other properties, like logP or MW as mentioned above.

Leave a Reply