Evolution on islands

A new study pub­lished in Evol­u­tion Let­ters has revealed how nat­ur­al selec­tion and genet­ic drift shape vari­ation in gen­om­ic and phen­o­typ­ic traits in island pop­u­la­tions. The study iden­ti­fied genes that could be influ­en­cing vari­ation in beak length and mal­aria infec­tion rates in pop­u­la­tions of Ber­th­el­ot’s pip­it Anthus ber­th­el­otii, a small pas­ser­ine bird that is found across island archipela­gos in the Atlantic Ocean. By look­ing at how these genes evolved, they showed that dif­fer­ences in beak length between archipela­gos appear to be driv­en by col­on­isa­tion his­tory rather than nat­ur­al selec­tion. In con­trast, vari­ation at genet­ic regions asso­ci­ated with mal­aria infec­tion sug­gests it is a rap­idly evolving trait. Here, first author Claire Arm­strong explains what her find­ings tell us about how island pop­u­la­tions evolve.

EL: What exactly is genet­ic drift, what effect can it have, and how does it dif­fer from selection?

CA: Genet­ic drift is an evol­u­tion­ary mech­an­ism that causes changes in the fre­quency of a genet­ic vari­ant (allele) with­in a pop­u­la­tion. Unlike selec­tion, where alleles will increase or decrease in fre­quency depend­ing on wheth­er they give an evol­u­tion­ary advant­age or dis­ad­vant­age, genet­ic drift changes allele fre­quen­cies at ran­dom. The smal­ler the pop­u­la­tion, the stronger the effects of genet­ic drift, so small, newly col­on­ised pop­u­la­tions can quickly end up with markedly dif­fer­ent gene pools to their ori­gin­al population.

EL: You stud­ied island pop­u­la­tions of Berthelot’s pip­it, a small song­bird. What made this study sys­tem par­tic­u­larly appro­pri­ate for your research?

CA: Islands are excel­lent places to study evol­u­tion. Pop­u­la­tions on dif­fer­ent islands will each exper­i­ence a dif­fer­ent set of selec­tion pres­sures, shaped by phys­ic­al dif­fer­ences between the islands, and by dif­fer­ent eco­lo­gic­al com­munit­ies formed by each island’s inde­pend­ent his­tory of col­on­isa­tions and extinc­tions. Isol­a­tion between islands means that indi­vidu­al island com­munit­ies often evolve inde­pend­ently, and as a res­ult are char­ac­ter­ised by unique flora and fauna.

Berthelot’s pip­its are found across three island archipela­gos in the Atlantic Ocean. Like many island spe­cies, these birds face an array of selec­tion pres­sures that may be caus­ing them to evolve loc­al adapt­a­tions, which over time could ulti­mately res­ult in spe­ci­ation between the island pop­u­la­tions. Berthelot’s pip­its in the north­ern archipelago of Madeira are clas­si­fied as a sep­ar­ate sub­spe­cies, based on longer beak lengths than in the more south­ern archipela­gos. We also find a large degree of vari­ation between islands in infec­tion rates for dis­eases such as avi­an mal­aria. By detect­ing genet­ic vari­ants asso­ci­ated with these traits, we can then ask ques­tions about the evol­u­tion of this spe­cies: are pat­terns of genet­ic vari­ation and trait vari­ation sug­gest­ive of nat­ur­al selec­tion, or are they more likely caused by ran­dom genet­ic drift?

EL: There are a huge vari­ety of dif­fer­ent traits you could have invest­ig­ated, includ­ing many related to repro­duc­tion and sur­viv­al. What is so spe­cial about beak length and mal­aria infec­tion? Are these traits import­ant for oth­er spe­cies as well? 

CA: We chose these two traits based on stud­ies by ourselves and oth­ers which showed that both beak length dis­ease res­ist­ance may be under diver­gent nat­ur­al selec­tion between the pip­it pop­u­la­tions. The evol­u­tion of bird beak shapes is very tightly linked with diet and feed­ing beha­viour, mak­ing this a very eco­lo­gic­ally import­ant trait. A change in diet can trig­ger rap­id evol­u­tion on beak shape, most fam­ously seen in Darwin’s finches when droughts have severely altered food avail­ab­il­ity. Infec­tious dis­eases such as avi­an mal­aria are also strong drivers of selec­tion. Avi­an mal­aria has had dev­ast­at­ing impacts on pop­u­la­tions of Hawaii­an birds, which have only recently come into con­tact with the dis­ease. In spe­cies that have had a longer co-evol­u­tion­ary his­tory with avi­an mal­aria, it can neg­at­ively impact body con­di­tion, lifespan, and the abil­ity to raise off­spring. We’re inter­ested in wheth­er dif­fer­ences in beak length and mal­aria infec­tion rates between pop­u­la­tions of Berthelot’s pip­its could be driv­en by nat­ur­al selec­tion. But, of course, these are just two traits and it’s cer­tainly true that oth­er traits will be diver­ging across pop­u­la­tions too.

EL: You used a range of genet­ic ana­lyses in this study, includ­ing RAD-seq and GWAS. Can you explain – for a non-spe­cial­ist – what each of these dif­fer­ent approaches tells you?

CA: RAD-seq, which stands for restric­tion site-asso­ci­ated DNA sequen­cing, is the meth­od we used to gain genet­ic inform­a­tion from our DNA samples. This approach uses restric­tion enzymes to chop up DNA into many frag­ments. We then sequenced thou­sands of these frag­ments scattered through­out the gen­ome in hun­dreds of birds. This approach of sequen­cing sev­er­al thou­sand genet­ic sites is more power­ful than sequen­cing just one or a hand­ful of genes, but still much cheap­er than sequen­cing whole genomes.

Once we had used RAD-seq to get our genet­ic inform­a­tion, we wanted to identi­fy genes that could influ­ence vari­ation in beak length and mal­aria infec­tion rates in Berthelot’s pip­its. We used a tech­nique called a gen­ome-wide asso­ci­ation study (GWAS), which looks at how well vari­ation at a genet­ic mark­er cor­rel­ates with vari­ation in a trait. Once we’ve found genet­ic mark­ers asso­ci­ated with our traits, we can look at where they are in the gen­ome, and see if there are any inter­est­ing genes close by.

EL: Once you had the res­ults back from your ana­lyses, what dif­fer­ences did you notice in terms of the effects of selec­tion and drift on your two traits of interest?

CA: Vari­ation in beak length and mal­aria infec­tion appear to be fol­low­ing very dif­fer­ent evol­u­tion­ary tra­ject­or­ies. We found a genet­ic mark­er that was asso­ci­ated with beak length with­in the Canary Islands, and when look­ing across the archipela­gos, we found the long-billed sub­spe­cies in Madeira only had the allele that was asso­ci­ated with longer bills. This makes com­plete sense, because birds on Madeira have longer bills in gen­er­al. When we looked at pat­terns of vari­ation at this mark­er across the islands, we found little evid­ence for nat­ur­al selec­tion at the DNA level. Togeth­er this indic­ates that vari­ation in beak length has likely been shaped by col­on­isa­tion events and genet­ic drift.

By con­trast, avi­an mal­aria pre­val­ence and vari­ation at asso­ci­ated genet­ic mark­ers did not reflect the col­on­isa­tion his­tory of Berthelot’s pip­its. Instead, pat­terns of asso­ci­ation var­ied between islands, without con­sist­ent diver­gence between the archipela­gos. This sug­gests that evol­u­tion of res­ist­ance is com­plex and con­stantly chan­ging, with dif­fer­ent pop­u­la­tions evolving dif­fer­ent res­ist­ance mechanisms.

EL: Do you have any idea how fast mal­aria res­ist­ance might be evolving in these populations?

CA: This is an area we’d like to get at in the future, by com­par­ing genet­ic vari­ation in museum spe­ci­mens with present-day samples. Our pre­vi­ous work has found that both mal­aria and immune gene diversity vary at the land­scape-level with­in a single island. This could sug­gest that evol­u­tion of mal­aria res­ist­ance is hap­pen­ing rap­idly, and at small spa­tial scales.

EL: Did your gen­ome scan high­light any oth­er traits likely to be cur­rently under diver­gent selec­tion between the archipelagos?

CA: We found a num­ber of genet­ic mark­ers that showed sig­na­tures of diver­gent selec­tion between the two sub­spe­cies. Look­ing at their pos­i­tions in the gen­ome, there were a couple that where situ­ated with­in genes that play a role in the immune sys­tem. Des­pite not find­ing evid­ence for diver­gent selec­tion for mal­aria at the archipelago-level, this could sug­gest there are oth­er aspects of the patho­gen com­munity that are exert­ing dif­fer­ent selec­tion pres­sures between archipela­gos. We also found diver­gent SNPs with­in genes that have links to meta­bol­ism; this could reflect cli­mat­ic dif­fer­ences between north­ern and south­ern Berthelot’s pip­it pop­u­la­tions, where lat­it­ud­in­al dif­fer­ences are driv­ing diver­gent selection.

EL: What’s next for your research?

CA: We’re really inter­ested in fol­low­ing up the immune and meta­bol­ism genes that we iden­ti­fied as show­ing sig­na­tures of nat­ur­al selec­tion. We’re also inter­ested in explor­ing vari­ation in our beak length and mal­aria genes, this time look­ing at how indi­vidu­als vary with­in islands to see if any land­scape factors drive vari­ation in genes, beaks and disease.

Claire Arm­strong (@ClaireArmstro) is a PhD stu­dent at The Uni­ver­sity of East Anglia. The study is freely avail­able to read and down­load here.