Testing the Predictability of Evolution

New research pub­lished in Evol­u­tion Let­ters uses a nat­ur­al exper­i­ment to gain insight into the roles of con­tin­gency and determ­in­ism in evol­u­tion. Sci­ence blog­ger Luke Turn­er tells us more.

Life has been evolving since it began on Earth, but just how pre­dict­able is the course of evol­u­tion? Anoth­er way to look at this ques­tion might be to ask; if we were to replay the tape of life and let evol­u­tion repeat itself, how sim­il­ar would the nat­ur­al world look to its cur­rent form? Some believe that it would adopt a com­pletely new format and likely lack humans. This school of thought was led by Steph­en J. Gould, who believed that con­tin­gency is the most import­ant part of evol­u­tion­ary the­ory. On the oth­er hand, people such as Simon Con­way Mor­ris argue that evol­u­tion is determ­in­ist­ic and there­fore pre­dict­able, with humans being inev­it­able products.

Debates between these two oppos­ing the­or­ies are ongo­ing, with evid­ence sug­gest­ing that both con­tin­gency and determ­in­ism have import­ant roles to play. One way to invest­ig­ate the rel­at­ive import­ance of these driv­ing forces is to study cases where dif­fer­ent lin­eages have col­on­ised the same envir­on­ment. If eco­lo­gic­al and genet­ic dif­fer­ences res­ult in spe­cies-spe­cif­ic evol­u­tion­ary out­comes, this would provide sup­port for con­tin­gency. In con­trast, if the shared selec­tion pres­sures of the same envir­on­ment lead to sim­il­ar char­ac­ter­ist­ics across spe­cies, this would indic­ate that evol­u­tion is more predictable.

Brand new research by Härer et al. (2018), pub­lished in Evol­u­tion Let­ters, attempts to resolve this debate, using an ideally suited nat­ur­al exper­i­ment. They stud­ied sev­en spe­cies of Nicara­guan cich­lid fishes, which con­cur­rently col­on­ised two types of envir­on­ment that dif­fer in their light con­di­tions. Around 500,000 years ago, cich­lids col­on­ised two great lakes in the Nicaragua Depres­sion, while much more recently – prob­ably with­in the last 2000 years – they col­on­ised vol­can­ic crater lakes from the great lakes. The envir­on­ments of the great lakes and vol­can­ic crater lakes dif­fer sig­ni­fic­antly in the amount of light that is avail­able. The great lakes are shal­low but murky, mean­ing that less light pen­et­rates through the water, while the crater lakes are deep but clear, allow­ing much more light through.

Top: Deep clear water of crater lake Xiloa. Bot­tom left: Murky, tur­bid water typ­ic­al of the great lakes and adja­cent rivers. Bot­tom right: Amma­t­it­lania siquia pic­tured in a clear-water crater lake. Pho­tos: Andreas Härer & Juli­an Torres-Dowdall

High vari­ab­il­ity in the visu­al sys­tem of cich­lids enables them to adapt their sight accord­ing to the amount of light in their sur­round­ings. This, coupled with the fact that their visu­al sys­tem is well under­stood, enabled the research­ers to study dif­fer­ences in the way they have adap­ted to their new envir­on­ments, and ulti­mately test how determ­in­ist­ic evol­u­tion is. Between the two light envir­on­ments, the research­ers found dif­fer­ences in how cich­lids expressed cone opsin genes, which are respons­ible for col­our vis­ion. Des­pite these genet­ic dif­fer­ences between spe­cies, the res­ult­ing changes in visu­al sens­it­iv­ity across spe­cies in the same envir­on­ment were con­ver­gent, sug­gest­ing these changes rep­res­ent adapt­a­tions to the light con­di­tions. Although dif­fer­ent spe­cies exhib­it clear dif­fer­ences in traits such as body size, col­our­a­tion and troph­ic level, all have evolved pat­terns of gene expres­sion that res­ult in sim­il­ar changes to their visu­al sens­it­iv­ity. This demon­strates that although there is extens­ive diversity between the spe­cies, com­mon selec­tion pres­sures act­ing on the visu­al sys­tem have led to con­ver­gent evolution.

Amphilo­phus cit­rinel­lus, one of the cich­lid spe­cies stud­ied by Härer et al. 2018. Photo: Claudi­us Kratochwil

While light con­di­tions appear to be driv­ing con­ver­gent evol­u­tion with­in the dif­fer­ent envir­on­ments, there were still dis­tinct dif­fer­ences in opsin cone expres­sion between spe­cies liv­ing in the same loc­a­tion. In oth­er words, even though dif­fer­ent spe­cies adap­ted to their light envir­on­ment in par­al­lel, cer­tain dif­fer­ences remained. For example, some spe­cies were more sens­it­ive to light than oth­ers, and these dif­fer­ences were main­tained across envir­on­ments. This indic­ates an import­ant sec­ond­ary role of con­tin­gency in cich­lid sight evol­u­tion, with eco­lo­gic­al dif­fer­ences between spe­cies also influ­en­cing their vision.

This research shows that although evol­u­tion may be unpre­dict­able over long time scales, it is pos­sible to pre­dict the course of evol­u­tion­ary change over short­er time peri­ods. Across the sev­en spe­cies of cich­lid fish stud­ied, con­ver­gent changes in the expres­sion of genes respons­ible for visu­al sens­it­iv­ity occurred when the fish moved to a new envir­on­ment, and these were pre­dicted by the amount of light avail­able in each loc­a­tion. Des­pite this, pre-exist­ing dif­fer­ences in cone opsin expres­sion between spe­cies were main­tained. These res­ults show that both determ­in­ism and con­tin­gency are import­ant in the evol­u­tion of cich­lid visu­al sys­tems, with con­ver­gent phen­o­typ­ic adapt­a­tion res­ult­ing from non-con­ver­gent changes at the molecu­lar level.

Luke Turn­er is a MSc Sci­ence Com­mu­nic­a­tion stu­dent at the Uni­ver­sity of Shef­field. The ori­gin­al study is freely avail­able to read and down­load from Evol­u­tion Let­ters here.