By David Lovell and Paul Berkman
Australia—like most countries—loves the warm inner glow that comes from “leading the pack”.
But off the sports field (or away from 1960s motorbike gangs) in more complex arenas like science, it’s less clear how to measure who is “leading the pack”.
Enter bibliometrics, which endeavours to measure the number and (ideally) impact of the science publication produced by individuals, institutions, … even nations.
When it comes to bioinformatics, this Hans Rosling-style visualisation gives us a bibliometrics insight into Australia’s pack-leading status. Please enjoy it (it’s interactive!) and for more context, read on…
Declaring our interests
We both care deeply about delivering positive impact from science, particularly through bioinformatics. We both live and have been educated in Australia
Our interest in the impact and standing of Australian bioinformatics stems not from a sporting concern for being “on top”, but from the commitment we share to do better, and deliver even more impact from science.
We sat up and took notice when our CSIRO colleague Anne Stevenson pointed out a study well outside our usual reading list
Productivity and Influence in Bioinformatics: A bibliometrics analysis
This study was published in the Journal of the Association for Information Science and Technology by Min Song, SuYeon Kim, Guo Zhang, Ying Ding and Tamy Chambers back in 2013.
But recent changes to higher education and research proposed by the Australian government make it timely to consider the study from an Australian context.
These changes include a substantial decrease in publicly funded research, notwithstanding an intent to establish a $20B medical research fund. And just as maths, stats, computer science and biology are fundamental to bioinformatics, so too bioinformatics is one of many disciplines fundamental to medical research.
With this in mind, we combined the publication information of Song et al., with OECD and IndexMundi data on population, GDP and research workforce to allow comparison of publication rates per capita, per $GDP and per researcher, over time.
Australian Bioinformatics: The leader of the pack?
The first question we explored was whether the number of bioinformatics publications produced by a country is in proportion to population. If that was the case, then each country would be producing the same number of publications per capita... and the bars in the graph on the right would all be much the same height.
Well done Switzerland!
Then we wondered about the number of bioinformatics publications produced per researcher (or full time equivalent - FTE) by country, and over time. We kept the same colour scheme as the preceding graph. Note that we are not able to specify particular kinds of researchers; this line chart shows publications per researcher in general. So we can't tell if the USA research workforce is more bioinformatics-oriented than other countries, or whether USA bioinformatics researchers pump out a whole lot more papers than their overseas colleagues.
But well done USA!
Finally, seeking "the full Rosling", we looked at publications per research FTE versus research FTEs per capita. This graph is dense with information and offers countless hours of interactive fiddling. We've highlighted the US, the UK and Australia and you can see that while the US and UK research workforce has stayed at the same proportion of population over the last decade, both countries have greatly increased their number of bioinformatics publications per researcher. Australia research workforce seems to have grown as a proportion of population, but its bioinformatics publication rate per researcher has increased at a more modest rate.
So, is Australian bioinformatics leading the pack? Well subject to the following caveats, we are yet to find a configuration of the data where Australian bioinformatics publication rates are far from the middle of the pack (and, as far as we know, there have been no songs ever written about that position).
OK, before you take these data too seriously, it is important to point out a few caveats.
- We are neither demographers, nor bibliometricians (is there even such a word?). For these and other reasons, we could regard the comparisons presented above as indicative rather than definitive.
- Each country shown on the charts have at most three independent data points about bioinformatics publications taken from Table 6 of Song et al.; Google charts linearly interpolates between them
- Note also that Table 6 lists only the top 20 countries, which means some countries don't appear, or appear only in some years
- We have approached Song et al. to see if we could get access to their data in more detail
- Publications and publication data provide only a limited window into research impact and researcher productivity. (For more on that, just ask anyone who's been knocked back by a journal recently.)
It is hard to draw simple conclusions from the data shown above. (Except for Switzerland... not just a land of great alpine beauty and fabulous cheese, but a hive of bioinformatic activity! You rock!)
So what are we trying to say in sharing this information?
- First, we want to raise awareness that bioinformatics is critical to modern bioscience, whether that be in medical research and health, agriculture, environmental science, human nutrition, resistance to antibiotics, biodiversity measurement, drug development, what have you...
- Second, and in relation to national aspirations in the life sciences, we want to encourage discussion as to whether each nation has the bioinformatics capability it needs to meet those aspirations.
- Third, we want to encourage people to look up and out, to ask what it is that makes some nations "leaders of the bioinformatics publishing pack" and to ask whether that's where they too want to be and, if so, how to get there.
We have to stop now...
We've got some papers to write.