2500 Galaxy Papers & Counting

August 12th 2015

Galaxy at CiteULike Social Bookmarking Service

The Galaxy CiteULike library recently reached a milestone: It now has over 2500 publications in it. The Galaxy CiteULike Group was launched in December 2011 and reached 1000 papers 18 months later.

To be included in the library a publication needs to reference or mention Galaxy, extend Galaxy, use or reference a Galaxy instance, or otherwise discuss Galaxy or cite one of the Galaxy Project papers.

Here's a review of those first 2500 papers.

The Tags

Each paper is reviewed and one or more tags are added to it. The initial set featured 9 tags. This wasn't quite enough, and 8 more were added in 2013, bringing the total to 17. (Papers from 2012 were back-curated with the new tags, but not before that.) The numbers for each tag in each year for the first 2500 papers are below.

Tag 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 #
methods 2 15 26 50 92 196 260 325 258 1224
workbench 1 3 7 11 18 36 68 127 151 205 121 748
usemain 1 1 3 89 95 64 253
tools 3 1 8 28 36 67 37 180
usepublic 1 15 57 77 150
isgalaxy 2 2 5 16 15 26 45 24 135
cloud 1 2 13 22 38 24 100
uselocal 2 28 41 27 98
shared 1 1 7 13 21 24 17 84
other 1 1 8 8 38 21 77
refpublic 10 28 26 64
reproducibility 6 7 8 24 12 57
unknown 2 5 7 3 10 13 8 5 53
project 1 2 1 5 6 10 6 7 9 47
howto 1 2 1 3 4 12 6 11 5 45
visualization 1 1 2 3 6 3 16
usecloud 2 1 1 4


A couple of trends stand out in the tag data:

  • methods and workbench have always been the most popular. methods papers use Galaxy in their analysis. workbench either just mention Galaxy or discuss the platform itself.
  • The numbers of usepublic and refpublic publications are climbing rapidly. Respectively, these are methods studies that did their analyses on a public Galaxy server other than usegalaxy.org, and papers that reference those servers in some way (besides their methodology). Part of this increase reflects better tracking of these papers, but (I believe) most of the increase reflects both the increased number of public servers (as reflected by the isgalaxy numbers), and their increased visibility.
  • reproducibility became a hot topic (finally) in 2014.


The Journals

Information on where the papers appeared is also available. Galaxy-related papers have appeared in over 500 different publications. Publications with 15 or more Galaxy papers are:

Journal 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Total
Plos One 0 0 0 1 1 4 16 32 42 45 27 168
Nucleic Acids Research 0 1 0 1 11 13 8 21 22 29 19 125
Bmc Genomics 0 0 0 0 2 4 12 22 27 42 15 124
Bioinformatics 0 0 0 5 4 7 23 19 19 30 10 117
Bmc Bioinformatics 0 0 1 0 2 10 15 16 8 25 6 83
Genome Research 1 2 4 3 5 2 9 12 2 8 3 51
Genome Biology 0 0 0 1 3 9 11 11 4 6 2 47
Plos Genet 0 0 0 2 0 1 6 10 15 6 4 44
Briefings in Bioinformatics 0 0 0 1 1 1 3 2 8 4 9 29
Genome Announcements 0 0 0 0 0 0 0 0 4 3 19 26
Concurrency and Computation: Practice and Experience 0 0 0 0 0 0 1 3 0 11 10 25
Proceedings of the National Academy of Sciences 0 0 0 1 1 0 2 3 4 8 4 23
Cell 0 0 0 0 0 1 1 5 8 3 5 23
Plos Comput Biol 0 0 2 1 0 2 3 3 5 5 2 23
Molecular Biology and Evolution 0 0 0 0 3 0 1 1 9 2 4 20
Nature Communications 0 0 0 0 0 0 0 2 4 7 7 20
Molecular Ecology 0 0 0 0 0 1 1 1 11 4 0 18
Database 0 0 0 0 0 1 6 2 4 3 1 17
Genome Biology and Evolution 0 0 0 0 0 1 3 2 3 5 2 16
Nature 1 0 0 0 1 3 1 0 4 2 3 15
Cell Reports 0 0 0 0 0 0 0 3 5 4 3 15


There are also many unexpected publications in the list:

  • Applied Stochastic Models in Business and Industry
  • Computers & Geosciences
  • Current Opinion in Solid State & Materials Science
  • Journal of Archaeological Science
  • ...

The Summary

Finally, the total number of papers per year continues to increase, and we expect 2015 to surpass 800 papers

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Total
Total 2 4 12 31 53 107 202 394 495 703 497 2500


We will continue to report new papers in the monthly Galaxy newsletters. New tags may also show up as the project and community evolve.

In the meantime, I expect the next 2500 papers will be published in considerably less than time than the first 2500. I'm looking forward to all that research

Dave Clements