Hello, I'm about to start a survey/literature review on "Protein Sequence Classification" or something of the sort. I was wondering if anyone has any advice on what tools I should be using. Particularly anything to keep track of all the references? searching for references? and what program to write it in?
Any and all comments are welcome
Comments
Have you tried Jabref?
Well it all depends if your are using a WYSIWYG editor, like word or openoffice or if you are using Latex. If you are using word, the easiest is porbably Endnote, but neither word nor endnote are free.
I am using Latex and Bibtex, and Emacs becomes then quite a powerfull tool to manage even large databases using flat files. I have nevertheless tried some of the freewares available and the simpliest and most useful one I have found is Jabref. It is java based, has a nice GUI, is free, works on any plateform, and is easy to install. It is also very straightforward to use, and has an import option from Web of Knowledge or Medline.
emacsen
I must admit I find it hard to get excited about latex through emacs. Yes, emacs is powerful, yes it can be incredibly useful (I use it for Sweave document preparation and R/ess sessions) but unless you speak elisp, it's hard to do cool stuff. And elisp is a little too esoteric for me.
I use Kile for latex, which allows customisable compiles, multiple-file projects, bib compiles, etc. Not perfect by any means, but very good for a minimal user investment.
Wish I Did
I've tried pybliographer, but I've had problems with trying to maintain a current list without duplicates. One solution involves some simple adjustments: you make regular (daily, weekly, etc) pybliographer queries of pubmed via command line (via cron) and then have a simple script parse out duplicates and load them into a MySQL database, which then can be manipulated as needed. Pybliographer comes with some tools, and working in between some different formats is feasable.
I like Endnote, I am currently trying to get it working via Crossover (Wine) on a SuSE box, so I can use it with MS Word (also via Crossover).
Or, finally, my current technique: a simple spreadsheet/flat file. I've found it works to keep it simple.
NCBI has simple e-tools for automatedly searching Pubmed. A couple of tools like PubCrawler, actually use e-tools for automated searches. BioMail provides a similar (free/opensource) service.
If you want to write something, I would suggest Perl as a language. Great for text parsing, etc, and lots of people have experience and built tools. (Python can't hurt either). You could use Perl's LWP protocol to do a simple Pubmed query, and then get results back in XML, and parse out the answers, and store as needed. It is _very_ simple and powerful. The work will come in creating a nice interface, and customizing the backend to your specifications.
Perl and Eutils
I also recommend Perl (of course!) and the NCBI EUtils services. This is exactly how I generate the "weekly publications" list over at ArchaeaWeb. Perl script runs as a cron job each Friday to access esearch/efetch using LWP and get the last 7 days publications in XML format. Then the Bioperl Bio::Biblio module is employed to reformat the XML to HTML and the HTML is uploaded to the website. You could also use an XML parser, or just regex for the tags, use RefDB to convert to RIS and import to MySQL...the possibilities are almost endless. So you need to spend some time deciding exactly what you want to do and designing the system accordingly.
Check out Alf's research tool tips
One of our contributors, Alf Eaton over at HubLog, has written a list of notes on tools for handling information, which you may find useful. Your query is not very specific - try tips for asking questions too.
Thanks
thats helpful, seems he recommends 'endnote' for what I'm looking for which would be called "Reference Management Software" I guess.. that
Pybliographer is good
I've tried Pybliographer - the search facility is nice, I find configuring the output is not very intuitive (admittedly I run debian unstable).
The OpenOffice people are working on draft specs for a built-in Endnote-like facility, which would be fantastic, but no code yet.
I have tried RefDB - it's a bugger to set up, but offers possibilities with MySQL and Perl bindings.
To be honest, literature management is an area where Linux/open source tools is lacking at the moment. I don't have any kind of system at all - most people that I know are stuck with Word/Endnote.
Reference databasing
Like all other answers, this one starts with "It depends..."
Specifically, on what setup you are currently using, how obsessive you want to be about searchable records, and what (if anything) you want to pay. There are two main options: EndNote-like commercial software, and markup technology tools for bibliographies.
The first option includes EndNote itself, Reference Manager, and ProCite (all essentially the same). These programs are reasonably easy-to-use, and have good interfaces to MSWord. However, they tend to be inflexible, and it's very hard to manage your local pdf copies through this (see note at bottom). I'm sure there are other alternatives, but these are the main ones I know of.
The second, more powerful (but harder/fussier) alternative is to use established markup technologies. The prime example is BibTex, which is a Latex-inspired markup for bibliographies. Rather obviously, it interfaces seamlessly with Latex, so if that's your document poison, you should go with BibTex. Hubmed provides .bib output. Pubmed itself produces XML output, which can either be converted to bib using several tools, or used straight with a home-grown parser or one of more than a dozen already written and incorporated into larger frameworks (DocBook, for example). You can database your records using tools such as RefDB or emacs bib modes.
The moral of the story is: if you use a commercial word-processor and have no (interest in aquiring) scripting skills, you're better off using the commercial solutions. If you want to have marked-up text records to search/db (look at swish-e), go with one of the markups. I prefer bibtex as I use latex for document preparation.
NOTE: to handle local pdf copies, put in a custom field set as a url and point it to the pdf (eg "C:\pdfs\Blogs_et_al.pdf", or"/home/me/pdf/Blogs.pdf").
NOTE: The markup technologies are plain text. Unless you're generating bibs with scripts, you probably want to use one of the countless GUI interfaces to bib records. Ditto XML.