Sign in or Join FriendFeed
FriendFeed is the easiest way to share online. Learn more »
Michael Kuhn
Quick poll for next STITCH version concerning chemical names
I need to shorten chemical names to show them in a protein--chemical network picture. The current algorithm creates abbreviations like "10-formyltetra." for 10-formyltetrahydrofolate, which might be better represented by "10-formyl-THF", which could be created by a new algorithm. However, there are edge cases: - Michael Kuhn
What would you prefer in these two cases: "acetyl-L-carni." or "ALCAR" for "acetyl-L-carnitine"? "androstanolone" or "dihydrotestost." for "dihydrotestosterone"? - Michael Kuhn
I am with Neil. Mouse over option would be great - Deepak Singh
May be a nice crowd sourcing exercise? - Egon Willighagen
yes, I guess I'll keep the current scheme for now ... problem is, my simple abbreviation generator fails for the >60000 chemicals that all start with "tert-butyl N-[" -- but most of them are not in the network anyway. - Michael Kuhn
Crowdsourcing won't work well with 10 million compounds (<100k of those will be in the network, though). I thought about using short names for the compounds that have been mentioned in PubMed (I guess this is similar to crowdsourcing), but as the above cases show, this is sometimes more confusing then a simple abbreviation of the name. - Michael Kuhn
if the target audience is composed of chemists, how about just showing the chemical structure as the node "name"? Full name and synonyms on mouse over. Also, I'm sure it only helps for a small number of examples, but PubChem has "Depositor-Supplied Synonyms". A third idea -- find the corresponding wikipedia page (by CAS/PubChem ID) and then look for other titles that redirect there (e.g., http://is.gd/1Kvhl). - Andrew Su
I already import all of PubChem, and based on PubMed text-mining I decide for a good name. If this doesn't work, I look at the databases mentioning a name and try to pick a good one. This generally works, but the problem is that I need to show a short name in the network, e.g.: http://stitch.embl.de/images... - Michael Kuhn
Yes, I was already looking at Wikipedia and there might be a way to get all the redirect / disambiguation pages that link to chemical names. The problem is that I might get names that are not yet in my list of synonyms, and I would want to avoid that. E.g. dihydrotestosterone I could get "5a-DHT" as a good short name from WP, but I don't get this synonym from PubChem -- so users couldn't find it in my database. - Michael Kuhn
Michael, I can imagine that crowdsource would certainly help for the places where that is needed... people won't care about the last 9.9 million... what you need is the name to be clickable so that people can suggest a *better* short name... then your crowdsourcing covers the subset for which it really matters... - Egon Willighagen