My first R package, phylosignal, that I started more than a decade ago, has just passed its 500th citation. So I decided to take a little time to celebrate (we definitely don’t celebrate enough). It’s a small milestone, but one that prompted me to look back on where the package came from, what it’s become, and what this experience has taught me about writing scientific software.

A folder turned into a package turned into a chapter

When looking back, it’s actually funny to remember that this package was never meant to be a big deal. It started as a side project and almost by accident during my PhD. At that time I was studying the phylogenetic signal in ecological preferences of freshwater diatoms. Working with phylogenetic methods in R involved a lot of scattered code that quickly started to accumulate in too many folders. Eventually, I decided to consolidate everything into a package. Partly, this was just to make my own life easier, but to be honest I also had a quiet dream of publishing my own R package.

At the time, I was already an avid user of R and yet this felt almost out of reach. R package development was still a bit of a dark art (at least that’s how I pictured it!). But very fortunately around that same time packages like devtools and Hadley Wickham’s book R Packages (1st ed.) came out, and suddenly, everything felt more accessible. That convergence of need and emerging resources was what, I think, made phylosignal possible. And that’s how, little by little, the package took shape. First as a collection of functions that I used and updated every day for my own PhD experiments, then as a real toolbox that could be used by others, to the point where it became a whole chapter of my PhD thesis.

A useful little thing

The paper introducing phylosignal was eventually published in Ecology and Evolution, and to this day, it remains my most cited work. What surprised me from the beginning was how widely the package was adopted. This is obviously a drop in the ocean compared with giants such as ape, adephylo, phytools, etc. but it’s been used in research well outside my field, cited by studies across a wide range of species and ecosystems. That breadth still gives me a sense of joy because there’s something very satisfying in knowing that this small piece of code you wrote can end up helping others explore completely different corners of biology.

Funny enough, the manuscript was first rejected by a different, more established, journal. The decision was motivated by a lack of novelty, and to be fair, this was not wrong. The package did not invent completely new methods. It mainly reimplemented well-known indices in C++ for speed, included some spatial autocorrelation tools in a phylogenetic context, and bundled it all with plotting and utility functions to make things easier. But I think the journal somehow missed something important: that scientific software is not just about methodological novelty.

As it turns out, the paper is now among the top 10 most cited in Ecology and Evolution (OpenAlex data). Of course, the number of citations is not an end in itself, but still it suggests that phylosignal has indeed filled a niche. That’s worth reflecting on because it shows that clarity and care in implementation, documentation, and API design, also matter. At the end it is heartening to see that a thoughtfully designed tool can have a long and successful life.

Still worth doing (I guess)

What is more frustrating, though, is how little this kind of contribution is recognized in the professional structures that shape our academic careers. Despite the clear and sustained need and use of research software, its importance in hiring, promotion, and evaluation is often invisible. I’ve seen this with phylosignal but also across multiple (possibly too many) projects I’ve invested in. It’s a bitter lesson, and one I keep relearning: that meaningful, widely used work can still fall through the cracks of what our institutions choose to value.

And yet, I keep doing it. I have often wondered why. Probably because there is still something deeply fulfilling about building things. This is especially true for something that lasts, something that others pick up and use in ways you never imagined. So here is to the next decade of phylosignal. With luck, care, and a bit of persistence, I might be writing another post for its 1000th citation.