If you're looking for a given ID, try this -- http://eutils.ncbi.nlm.nih.gov/entrez... -- then use grep to extract all lines which start with MH to get the MeSH keywords - Andrew Clegg

Obviously you need something like curl or wget which can run from a script and download from a URL. If you need to search by something other than PMID (e.g. a text query), use esearch to run the query and get back the IDs, then efetch to download the docs corresponding with those IDs. Does that make sense? - Andrew Clegg

e.g. to get all MeSH terms and the pub date for PMID 19614587, try this (from a Mac or Linux command line anyway): curl -s 'http://eutils.ncbi.nlm.nih.gov/entrez... '^(MH |DP )' -- ARGH, FriendFeed being 'helpful' has mangled the command and I don't know how to override that. If you give me an email I can send better instructions :-) - Andrew Clegg

Download you data as XML ( http://eutils.ncbi.nlm.nih.gov/entrez... ) and then extract the MESH terms using XSLT. Do you need help for this ? For example, see the http://code.google.com/p... in my repository. - Pierre Lindenbaum

Pierre, mere mortals can't write XSLT. ;) At least I failed miserable the last time I tried to do something simple - Michael Kuhn

tutututu, XSLT is simple for simple things (http://www.w3schools.com/xsl... ) . and XSLT is complicated for difficult things ( http://docbook.sourceforge.net/release... ). I mean , it takes less than one hour to make a simple task with XSLT. - Pierre Lindenbaum

and the more people wil use XML+XSLT or RDF, and the more you will work with structured data instead of parsing/grepping/sorting meaningless tab-delimited files. - Pierre Lindenbaum

Yes, it's only starts to become fun when you are pulling XML from more than one document :) - Egon Willighagen

personally, I would use an xpath library to parse the xml and output in plain text. my understanding is that xslt (which uses xpath) is great for parsing xml but pretty much is designed to output XML or HTML only. - Andrew Su

@andrew : can you use Xpath without programming ? which of tool is it ? Anybody can use XSLT as most browser embed an internal XSLT engine. You just have to add the URL of your stylesheet at the top of your XML document ( http://www.w3schools.com/xsl... ) - Pierre Lindenbaum

Seriously guys -- XPath/XSLT is overkill for this kind of thing. It's a one-liner. I've written is up here so it doesn't get mangled this time: http://biotext.org.uk/how-to-... - Andrew Clegg

working code wins! ;) (But I will say that learning to use the perl XPATH library was surprisingly simple, even for someone whose skills are decaying as fast as mine...) - Andrew Su

@Andrew C.. yes of course , I agree :-) - Pierre Lindenbaum

BTW XMLStarlet is great for doing XPath/XSLT/etc. from the command line -- http://xmlstar.sourceforge.net/ or you may already have it as /usr/bin/xml -- but try it the easy way first ;-) - Andrew Clegg

Andrew C, where I think your one-liner gets a bit more complicated is when you request multiple documents (e.g., http://eutils.ncbi.nlm.nih.gov/entrez...) and then want to parse out multiple fields from each article on a single line. That I think is where the balance shifts a bit toward xpath... - Andrew Su

@Andrew Yeah, it very much depends on what you want to do *next*... - Andrew Clegg from email

recently, @tpoi asked for a tool returning the number of paper per year ( http://twitter.com/tpoi... ). I wrote this stylesheet http://code.google.com/p... . He was then able to compute the result without knowing any command-line tool, but using his browser. - Pierre Lindenbaum

(sorry to everyone else for hijacking this thread, but...) Pierre, very cool. How do you invoke your xsl? you need to reference it in an XML doc, right? - Andrew Su

1) you can invoke a XSLT engine e.g. http://xmlsoft.org/XSLT... "xsltproc style.xsl doc.xml" 2) you can add a stylesheet url at the top of your XML document and open this document in firefox . See "Link the XSL Style Sheet to the XML Document" in http://www.w3schools.com/xsl... - Pierre Lindenbaum

@All : Thanks a lot for the awesome responses. Let me go through them and try them. I was working on to contribute to GeneWiki3 (Thanks to Andrew Su, he assigned a small module for me to contribute to GeneWiki3) This is a true Science 2.0 learning experience. Thank you all !! - Khader Shameer