Discussion

John and Joshua, I’d like to make a few points about NIH/CADD Resolver that may be helpful. First, Resolver is not really a chemical data source. It’s primary function is the interconversion of chemical identifiers. I’ll reference the identifiers name, IUPAC, InChI, InChIKey, CAS, SDF (2D and 3D), and SMILES in this note. 1) Name Identifier: This is the weakest identifier of all, and it is Resolver’s weak link. A given chemical compound can have a huge number of names. When Resolver is presented with an input name, it must use look-up. IUPAC is a different matter, but I’ll get to that in a minute. Bottom Line at Resolver: Look-up Weak. 2) CAS Identifier: This is almost as bad as name! Many compounds have more than one CAS number. Here are CAS numbers for ethanol from Resolver: 121182-78-3 64-17-5 8024-45-1 8000-16-6 68475-56-9 71076-86-3 71329-38-9 There are two reasons for this multiplicity: Chemical Formulations and Money (CAS sells these things!). Bottom Line at Resolver: Look-up Weak. 3) IUPAC, InChI, SMILES, SDF Identifiers: These identifiers are all exact codes for the structural atom connectivity of a molecule. OChem Textbooks, of course, cover IUPAC, but IUPAC is only one member of this powerful group. With any of these as input, Resolver uses powerful algorithms to convert input to any of the others that may be need. No Look-up Involved. What the heck is this all about? All databases must store information with reference keys. Not all databases speak the same key language. Resolver’s job is to take the identifier that you have and convert it to the identifier that you need. Bottom Line at Resolver: NO Look-up Strong. For example, the very popular Web structure drawing application JME (also JSME) creates SMILES. NIST allows spectrum lookup by InChI. But if you want to get the actual spectrum at NIST, you also need the CAS number that NIST uses for that particular compound. Yikes! Fortunately, NIST will give it to you if you ask nicely using the InChI - it turns out that they usually use the lowest CAS number. By the way, Google indexes InChI and IUPAC. InChIKey Identifier: We’re back to look-up, but this is look-up with power. InChI’s tend to be long. InChIKey’s are true calculated hashes of InChIKey. They are darn close to being truly unique. Unique is what you want in look-up. Google also indexes InChIKey. So your spread sheet “database” exercise uses Resolver because it’s easy to use, but it’s just a warm up for what’s to come. What might have been lost is this initial warm up exercise is that you were using a tool that is really not a database (Resolver). Rather it’s one of the keys for opening real databases. Otis