When we first launched "Nature Protocols ":http://www.nature.com/nprot I was keen to get our protocols indexed by all the big players. Whilst there was going to be a marketing campaign with ads in Nature and banner ads all over nature.com I knew we’d only get real users when people did their literature search and one of our protocols came top of the list. We wanted to be something researchers needed and found useful. However, alongside getting listed in the ISI Web of Knowledge came the news we’d also be eligible for an Impact Factor. It was clearly something scientists cared about. A lot of our early correspondence was about what our Impact Factor was. We had to be very apologetic and explain that we didn’t have one – first we’d need to publish some protocols, and then wait to see if they were cited. Personally I was relieved we didn’t have one because I wasn’t sure what an impact factor would mean for us, even if we did have one.

The Impact Factor comes into its own as a comparative factor. That’s why my colleagues upstairs at Nature are delighted with theirs, as at 36.101 it puts them ahead of their competitors. And also why journals with impact factors of 1 or 2 can also be delighted, as such an Impact Factor can be the highest in their specialist field. But when I looked into Impact Factors I realised that no protocols publications had them, so as a comparative factor, our impact factor would be rather meaningless. When and why would people cite protocols? We certainly wouldn’t be cited because we presented exciting new data and reported a major advance in scientific knowledge. I hoped that we’d be cited if we were used – and so lots of use would lead to a high impact factor. We’d obviously only get cited if the experiment worked, and so a higher impact factor would indicate good, reliable protocols. But, maybe we wouldn’t get cited, as citing the source of your protocol didn’t seem to be the ‘done thing’ at the time.

Five years down the line – with a 2 year ISI Impact Factor of 8.362 – I am pleased to report that we are being cited. We have an Immediacy Index of 1.547 and "Article Influence Score ":http://en.wikipedia.org/wiki/Article_influence_score of 3.799. We’re all rather pleased, as we know our Impact Factor matters to some of our authors. We’re also curious – which articles have people been citing? Are they the methods people have been using? Have they been getting good results? How do these articles correlate with the articles that are most downloaded from our site?

Our most highly cited article according to ISI is Locating proteins in the cell using TargetP, SignalP and related tools by Emanuelsson et al. This is different from the most highly cited article according to Scopus – "Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources ":http://dx.doi.org/10.1038/nprot.2008.211 – by Huang et al., although this comes in at second place. At third we have "Analyzing real-time PCR data by the comparative C-T method ":http://dx.doi.org/10.1038/nprot.2008.73 by Schmittgen et al. And in fourth, "In-gel digestion for mass spectrometric characterization of proteins and proteomes ":http://dx.doi.org/10.1038/nprot.2006.468 by Shevchenko _et al. _and in fifth, Protein structure prediction on the Web: a case study using the Phyre server. Interestingly there are lots of bioinformatics protocols in this list.

Of course, our 2010 2 year Impact Factor is only based on articles published in 2008 and 2009, so only two of those protocols, those in second and third place, will have been included. And also, only the cites in 2010. I was – and remain – curious about our five year impact factor. Whilst methods are developing all the time, protocols do tend to have a long life. A tweak here and there (hopefully highlighted as a comment to the protocol), but we hope our more useful protocols will still be in use, and cited, 5 years on. I guess we’ll have to wait and see