Lets publish the Wiki!

[dan]

Originally Posted by lh3

1) I am sure my paper has been cited more than "-1" times. But I do not know how to update the information. Sorry.

2) If permitted (by FindIt and/or web of knowledge), it would be good to provide a link like this (note how this URL links to PMID). This will give us the "official" citation numbers without a web of knowledge account.

Are you sure you don't have -1 citations? ;-) I'm working to improve this section, it's a little tricky to get the data, but now the memory problems on the wiki are fixed (I hope), we can start filling in these values again.

Ah, I see the citation count on findit now. Well... If you can build an API interface to a findit screen scraping script, I'd be happy to use that in SEQwiki. I'd need something like:

http://some.box/findit.cgi?q=<PMID|TITLE>


It could return JSON, XML or CSV, as described here: http://www.mediawiki.org/wiki/Extension:External_Data


You could host it on the Google app server if you like python.


Cheers,
Dan.

[marcowanger]

On the right panel, you see

More like this

258 Articles that cite this
from Scopus
237 Articles that cite this
from Web of Science®

[lh3]

It seems that FindIt does not hold the data. When FindIt displays a page, it queries WebOfScience (WOS) on the fly. WOS must provide an interface to get the data. Nonetheless, WOS is a paid site. I do not know if the interface is public.

Another citation resource you may consider is <http://www.ebi.ac.uk/citexplore/webservice.jsp> from EBI. Still WOS has a better coverage, though.

PS: As I have just checked, CiteXplore also links to WOS, so there must be some way to query WOS. <example: http://www.ebi.ac.uk/citexplore/citationDetails.do?externalId=19451168&dataSource=MED&citedCount=1>

[marcowanger]

It seems you have to be licensed to use WOS api

http://bibwild.wordpress.com/2009/04...link-resolver/

[lh3]

Is it legitimate to get the WOS citation number from this CiteXplore link? It is just a plain HTML page, fairly easy to parse out.

[dan]

Originally Posted by lh3

Is it legitimate to get the WOS citation number from this CiteXplore link? It is just a plain HTML page, fairly easy to parse out.

My personal opinion is, if data is available on a website without registration, it's legitimate to use it (with suitable acknowledgements where appropriate).

This is the core idea behind the creation of the world wide web.

I think people who try to protect data 'on the web' have forgotten what the web is.

If you don't want it downloaded, don't put it on the web.

Just my 2p.

[marcowanger]

Dear Dan,

Quickly wrote something that fetch abstract and citation count based on PMID.

May you modify it to fits your mediawiki?

Let me know what I can help.

p.s. Thanks Heng for the EBI's link.

Regards,
Marco

Originally Posted by dan

Are you sure you don't have -1 citations? ;-) I'm working to improve this section, it's a little tricky to get the data, but now the memory problems on the wiki are fixed (I hope), we can start filling in these values again.

Ah, I see the citation count on findit now. Well... If you can build an API interface to a findit screen scraping script, I'd be happy to use that in SEQwiki. I'd need something like:

http://some.box/findit.cgi?q=<PMID|TITLE>


It could return JSON, XML or CSV, as described here: http://www.mediawiki.org/wiki/Extension:External_Data


You could host it on the Google app server if you like python.


Cheers,
Dan.

[lh3]

I am a little concerned about the list of "most popular tools" at the front page of wiki. I think it is hard to find the right criteria about what should go into the list. For example, should we include EMBOSS suite, BLAST, FASTA, BioPerl/BioPython and CAP3? CAP3 should certainly go to the list. We may still use BLAST/FASTA to map reads in highly diverged regions (at least I use BLAST occasionally). I know people are using EMBOSS to convert formats. BioPerl/BioPython are certainly widely used. These tools all have at least a couple of thousands of citations and will immediately take the top few places in the list.

On the other hand, the citations for the Erange and ChIPSeq software are counted based on the papers describing the biotechnology rather than on the methods. I cite the two papers when I refer to chip-seq and rna-seq in general. But I do not cite them for the methods. The ChIPSeq software is rarely used so far as I know (the link is broken now). Erange is still widely used, but I am not sure if it is the most popular any more.

The most accurate measurement of popularity of a tool is the number of citations for using the tool to process sequencing data. For some papers (e.g. maq/bowtie), all citations are like this. But for some other papers (e.g. BLAT/MUMmer/erange/chipseq), we can hardly know the answer. If we do not have solid criteria about what should be included in the frontpage, my opinion is to drop the list. Showing misleading data might be worse than showing nothing. Perhaps giving the timeline of each publication may be helpful in some aspects, but I guess it is very difficult to get this information and does not really solve all the problems.

Anyway, whatever decision you made, I really appreciate all your efforts. Collecting citations is a hard thing.

[dan]

Can you post this on a new thread?

I think it's worth discussing, but not here.

Cheers,
Dan.