Shaping evolution: selection versus constraint

In our new author blog, Joel McGloth­lin explains how nat­ur­al selec­tion and genet­ic con­straint inter­act to shape adapt­ive radiations.

Ever since the Mod­ern Syn­thes­is of the early 20^th cen­tury, bio­lo­gists have had a pretty good under­stand­ing of how evol­u­tion works. Muta­tion gen­er­ates genet­ic vari­ation in a pop­u­la­tion, and nat­ur­al selec­tion sorts through it, keep­ing things that fit the cur­rent envir­on­ment and weed­ing out things that don’t. Along the way, pop­u­la­tions some­times split into two, adapt­ing to dif­fer­ent envir­on­ments and becom­ing new spe­cies in the process.

All evol­u­tion­ary bio­lo­gists agree about these basics of the evol­u­tion­ary pro­cess, but there is still a lot of con­tro­versy about the rel­at­ive import­ance of its dif­fer­ent aspects. For example, some sci­ent­ists place more weight on the import­ance of nat­ur­al selec­tion as an explan­a­tion for bio­lo­gic­al diversity, while oth­ers think that genet­ic vari­ation plays a guid­ing role in con­strain­ing which paths adapt­a­tion can take.

It was with this big ques­tion in mind that my col­leagues and I star­ted on a large-scale quant­it­at­ive genet­ic study of Anolis liz­ards. Anoles are icons of what is known as adapt­ive radi­ation, an evol­u­tion­ary pat­tern involving the ori­gin of group of related spe­cies that dif­fer in both their appear­ance and eco­lo­gic­al role. In our paper, pub­lished today in Evol­u­tion Let­ters, we describe our efforts to dis­en­tangle the roles of nat­ur­al selec­tion and genet­ic con­straints in gen­er­at­ing this diversity of form.

My co-authors (Butch Brod­ie, Jonath­an Losos, Megan Kobiela, Helen Wright, Luke Mahler, and Jason Kolbe) and I estim­ated pat­terns of genet­ic vari­ation under­ly­ing eco­lo­gic­ally import­ant traits (called “G” for short) in sev­en dif­fer­ent spe­cies of anoles. We cap­tured hun­dreds of adult liz­ards from three islands (Puerto Rico, Jamaica, and South Bimini, Bahamas) and returned them to the lab, where we bred thou­sands of babies over the course of sev­er­al years. The breed­ing effort was massive and was only pos­sible with the help of a huge team of stu­dents and tech­ni­cians work­ing every­day in the anim­al facil­ity. Every one of these baby liz­ards was X‑rayed so we could meas­ure skelet­al traits like head shape and limb length to gen­er­ate an estim­ate of for G for each species.

These meas­ure­ments of G allowed us to test a num­ber of hypo­theses about genet­ic vari­ation and its influ­ence on the evol­u­tion­ary pro­cess. First, we asked wheth­er the evol­u­tion­ary dif­fer­en­ti­ation of anole mor­pho­logy was related to genet­ic vari­ation described by G. A clas­sic pre­dic­tion in evol­u­tion­ary quant­it­at­ive genet­ics is that spe­cies should diverge along the “genet­ic line of least of res­ist­ance,” or “g_max,­­‚” defined as the com­bin­a­tion of traits with the most genet­ic vari­ation, for at least a few mil­lion years. In anoles, we found that diver­gence of mor­pho­logy was indeed biased in the dir­ec­tion of g_max, even though the spe­cies had been sep­ar­ated for up to 40 mil­lion years.

This res­ult was sur­pris­ing to us, because the rela­tion­ship between g_max and diver­gence is pre­dicted to decay over time. One reas­on for this decay is that G itself is expec­ted to change as spe­cies evolve, which might obscure the rela­tion­ship by chan­ging g_max. Because we meas­ured G in mul­tiple spe­cies, we were able to meas­ure this change. Our res­ults show that G changes sig­ni­fic­antly in size, shape, and ori­ent­a­tion. Why, then, did the rela­tion­ship between G and diver­gence persist?

Using a meth­od called genet­ic cov­ari­ance tensor ana­lys­is, we broke change in G down into its com­pon­ent parts. This showed us that although G changed, the way it changed did not alter the genet­ic lines of least res­ist­ance. In fact, most of the change in G involved grow­ing and shrink­ing in the dir­ec­tion of g_max.

Put­ting all these res­ults togeth­er, we found that three axes—genetic vari­ation with­in spe­cies (g_max), change in mor­pho­logy across spe­cies, and change in genet­ic vari­ation across species—were all aligned with one anoth­er. This res­ult is open to a wide vari­ety of inter­pret­a­tions, which we dis­cuss fur­ther in the paper. At one extreme, genet­ic con­straint could be the whole story. Spe­cies are con­strained to evolve along genet­ic lines of least res­ist­ance, and even though these con­straints evolve, the con­straints them­selves are con­strained to evolve in cer­tain dir­ec­tions. At the oth­er extreme, selec­tion reigns supreme, and diver­gence resembles g_max because they have both been shaped by adapt­a­tion. The real­ity is prob­ably some­where in between.

When we began this study, we thought we might be able to show that either selec­tion or genet­ic con­straint determ­ined the paths of adapt­ive radi­ation, but instead, we may have demon­strated just how dif­fi­cult these forces are to sep­ar­ate. At least in anoles, con­straint shapes the evol­u­tion­ary response to selec­tion, but also evolves in response to selec­tion in such a way to keep the two entwined. Per­haps it’s this nev­er-end­ing cre­at­ive dance that makes evol­u­tion so inter­est­ing in the first place.

Joel McGloth­lin is Assist­ant Pro­fess­or in the Depart­ment of Bio­lo­gic­al Sci­ences, Vir­gin­ia Tech. The study is freely avail­able to read and down­load here.