MRE Symonds, SP Blomberg (2014) A primer on phylogenetic generalised least squares (PGLS). In: Modern Phylogenetic Comparative Methods and Their Application in Evolutionary Biology: Concepts and Practice (ed. LZ Garamszegi), Chapter 5, pp 105-130. Springer, Berlin. (unfortunately for copyright reasons, I cannot make this pdf available on the website, you are welcome, though, to contact me for a copy).
What’s it about?
PGLS is one of the principal phylogenetic comparative methods used in evolutionary biology. It’s main use is to be able to control for phylogenetic relationships when doing analysis of relationships between biological traits. However, there hasn’t not been a really good beginners level guide that explains what PGLS is and how it works (and also points out where you can go wrong). This paper, part of a new volume on phylogenetic comparative methods designed in part to replace Harvey & Pagel’s classic 1991 tome on the comparative method, is designed as that basic introduction.
What’s the story behind it?
I’ve had contact for several years with László Garamszegi, due to my participation in some statistical workshops that he organised at the International Society for Behavioural Ecology conferences in 2006 and 2008. At his instigation I contributed (with Adnan Moussalli) a paper to a special edition of Behavioral Ecology and Sociobiology that served as a beginner’s guide to using Akaike’s Information Criterion. Ultimately that paper was a very successful and fruitful contribution (it’s the journal’s most-cited paper from the last 5 years), although it’s fair to say the process of putting it together was agony, in part thanks to 6 extremely picky reviewers. Presumably it was with the success of this earlier primer in mind that Las contacted me in early 2013 and asked me to contribute a similar chapter to a new book on phylogenetic comparative methods that he was editing. My initial inclination was actually to refuse, because the pain and effort that the first paper had caused me didn’t seem like something I’d want to repeat. However, two things persuaded me to have a try 1) the other authors in the book are the top people in the field (I resisted the temptation to indulge myself too much in impostor syndrome here), and 2) I had been frustrated for a long time that there was not a good beginners level intro to PGLS in the literature – which given its widespread use seemed like a bit gap in the literature I could fill.
Anyway, I agreed, but then thought I’d better get a real expert on board and contacted Simon and asked him to help me make sure I didn’t make a fool of myself. (I’d met Simon years earlier when I went to give a seminar at Australian National University when he was there and he sat opposite me at dinner at some strange pizza restaurant out in woop-woop). It turned out my inclination to avoid pain was well founded, because one of the reviewers HATED the chapter. In a review that, to me, was borderline unprofessional in its tone, the reviewer dripped sarcasm and humiliation on everything (and boy was there a lot) that was wrong with the chapter. As with any of those types of reviews you have to put it aside to allow the anger to subside (a few weeks in my case) and then try and work out ‘ok, what are the points that they make that I should actually take on board’. After a substantial revision, the paper got accepted. I actually think the book should be a fantastic contribution to your book shelf, if you have any interest in phylogenetic comparative approaches. However, I have decided absolutely not to write one of these bloody introductory statistical methods papers again though…. until the next time I’m asked that is.
The Symonds Lab 