Environmental DNA: how to track biodiversity “barcodes”

Biod­iversity is in the midst of a crisis¹ and it is both import­ant to mon­it­or the evol­u­tion of wild pop­u­la­tions and to dis­turb them as little as pos­sible. There two seem­ingly con­tra­dict­ory actions could be recon­ciled thanks to an increas­ingly use­ful source of inform­a­tion: envir­on­ment­al DNA (or eDNA).

Since the 1980s, samples col­lec­ted from the envir­on­ment have made it pos­sible to study the invis­ible micro-organ­isms they con­tain by ana­lys­ing their gen­omes. This approach, developed thanks to advances in sequen­cing tech­niques and com­puter tools, gave rise to meta­ge­n­om­ics, the large-scale study of gen­omes.

But although micro-organ­isms are to be found just about every­where, they are not the only sources of DNA in the envir­on­ment! Mucus, dead skin, hair, car­casses, fae­ces… All liv­ing beings, whatever their size, leave traces of their pas­sage. And these traces can con­tain DNA that makes it pos­sible to track their pres­ence, which was demon­strated for the first time in 2008 by a team from the Alpine Eco­logy Labor­at­ory, which iden­ti­fied the DNA of bull­frogs in vari­ous ponds².

eDNA is there­fore also a tool for mon­it­or­ing pop­u­la­tions of macro-organ­isms, whose effect­ive­ness has been demon­strated in many envir­on­ments, from sed­i­ments to the seabed… and even in the air itself! Two stud­ies pub­lished at the begin­ning of 2022,³⁴ have shown that by suck­ing air from zoos and ana­lys­ing its DNA con­tent, it is pos­sible to identi­fy dozens of spe­cies of anim­als liv­ing in or near these zoos. This prom­ising approach opens up new pos­sib­il­it­ies but still has a num­ber of weak­nesses. To under­stand them, we need to look in more detail at how it works.

Once recovered from the envir­on­ment, DNA can be ana­lysed in three ways. The first is to per­form glob­al meta­ge­n­om­ics, sequen­cing all the DNA to try to identi­fy as many of the gen­omes present as pos­sible. This approach is suit­able for the study of micro-organ­isms, which are dir­ectly con­tained in the sample and whose DNA is there­fore well pre­served and present in large quant­it­ies. It is less rel­ev­ant for macro-organ­isms, whose DNA is rarer and more dam­aged in the envir­on­ment, gen­er­ally being found in the form of frag­ments of a few dozen to a few thou­sand nuc­le­otides long.

The ana­lys­is of this type of eDNA relies on the use of molecu­lar bar­codes, genet­ic sequences that identi­fy a spe­cies or class of organ­isms. These must be short enough to be found in nat­ur­ally frag­men­ted DNA. These “bar­code” sequences are spe­cific­ally amp­li­fied by PCR and then sequenced and ana­lysed. They can be more or less spe­cif­ic, mak­ing it pos­sible to mon­it­or a single spe­cies, a group of closely related spe­cies or to make a more extens­ive invent­ory of the biod­iversity of an envir­on­ment. Depend­ing on the range chosen, this is known as bar­cod­ing or metabar­cod­ing, which is a form of tar­geted metagenomics.

The ana­lys­is is then based on the com­par­is­on of the bar­codes obtained with those lis­ted in data­bases. These data­bases are more or less well pop­u­lated depend­ing on the field and this is one of the weak points of eDNA stud­ies: the rich­er the data­bases, the less lim­ited the ana­lyses. The accu­mu­la­tion of new data is gradu­ally chan­ging the situation!

Cur­rently, eDNA has four main types of applic­a­tion: the study of biod­iversity (cata­loguing of spe­cies, mon­it­or­ing over time, ana­lys­is of bio­lo­gic­al func­tions⁵, etc.); the tar­geted mon­it­or­ing of cer­tain spe­cies (par­tic­u­larly threatened, invas­ive or bioin­dic­at­or spe­cies⁶); the estim­a­tion of the rel­at­ive abund­ance of spe­cies in a giv­en envir­on­ment; and the recon­struc­tion of diets by ana­lys­ing the DNA con­tained in excre­ment⁷.

In some envir­on­ments, such as sed­i­ments and very cold envir­on­ments, eDNA is pre­served for long peri­ods of time, mak­ing it pos­sible to go back in time. For example, research­ers have stud­ied uni­cel­lu­lars from the Brest har­bour over a peri­od of 1,400 years, high­light­ing the impact of the Second World War and recent changes in agri­cul­tur­al prac­tices⁸. The record for the use of eDNA in palaeoe­co­logy was broken at the begin­ning of Decem­ber thanks to per­ma­frost samples from Green­land, which made it pos­sible to recon­struct a palaeoe­cosys­tem of about two mil­lion years⁹!

eDNA also has many advant­ages for the study of present-day spe­cies. Indeed, its col­lec­tion is non-invas­ive, which avoids dis­turb­ing the envir­on­ments stud­ied, and very simple. The cor­res­pond­ing field­work requires little equip­ment and train­ing, allows access to places unsuit­able for dir­ect obser­va­tions, can be eas­ily graf­ted onto exped­i­tions already planned else­where and remains decoupled from the ana­lys­is work. It is much less costly and restrict­ive than con­ven­tion­al obser­va­tion meth­ods, allows for more sampling and opens up pos­sib­il­it­ies for mon­it­or­ing on a large spa­ti­otem­por­al scale. Ana­lyt­ic­al meth­ods also lend them­selves to this expan­sion, as pool­ing the pro­cessing of many samples gen­er­ates eco­nom­ies of scale.

In some envir­on­ments, such as sed­i­ments and very cold envir­on­ments, eDNA is pre­served for long peri­ods of time, mak­ing it pos­sible to go back in time.

As prom­ising as it is, how­ever, the use of eDNA has its lim­it­a­tions. To begin with, and this is fun­da­ment­al even if it seems obvi­ous, detect­ing an indi­vidu­al’s DNA is not the same as detect­ing their pres­ence. It does not tell us any­thing about its state of health, size, or stage of devel­op­ment, all of which are only access­ible through dir­ect obser­va­tion. It does not neces­sar­ily tell us where it is, either, as DNA can be trans­por­ted in the envir­on­ment! In rivers, spe­cies can leave traces sev­er­al kilo­metres down­stream from their actu­al position.

Moreover, not all organ­isms release DNA in a com­par­able way into their envir­on­ment and, depend­ing on the fra­gil­ity of the struc­ture con­tain­ing the DNA and the con­di­tions of the envir­on­ment (in par­tic­u­lar pH and tem­per­at­ure), eDNA can be degraded more or less rap­idly. The absence of DNA does not neces­sar­ily mean the absence of a spe­cies. Con­versely, DNA can eas­ily con­tam­in­ate samples, wheth­er it comes from the exper­i­menters and their equip­ment or from anoth­er source. For example, res­taur­ants or mar­kets in coastal areas can lead to the detec­tion of eDNA from non-liv­ing fish¹⁰.

Finally, bio­molecu­lar and com­puter pro­cessing tech­niques gen­er­ate their own biases. Bey­ond the incom­plete­ness of the data­bases, PCR amp­li­fic­a­tion is not equally effi­cient on all DNA sequences and, depend­ing on the bar­codes chosen and the way they are detec­ted, false neg­at­ives or false pos­it­ives may appear, which is dif­fi­cult to mon­it­or on a case-by-case basis in large-scale ana­lyses. This vari­ab­il­ity in amp­li­fic­a­tion, togeth­er with sampling bias, lim­its the suit­ab­il­ity of eDNA as a quan­ti­fic­a­tion tool. Each study of this type requires metic­u­lous checks to ensure that the res­ults obtained via eDNA are com­par­able to those obtained by manu­al count­ing. The final icing on the cake of com­plic­a­tions is that the sequen­cing itself can be a source of error.

Solv­ing these tech­nic­al prob­lems is a cent­ral issue for teams inter­ested in eDNA. Sev­er­al research pro­jects aim to reduce the uncer­tain­ties of ana­lys­is, not­ably by stand­ard­ising pro­to­cols, pop­u­lat­ing data­bases and identi­fy­ing rel­ev­ant molecu­lar bar­codes¹², which bodes well for future improvements.

As Sam Chew Chin, a PhD stu­dent study­ing fish pop­u­la­tions via eDNA, puts it, this tool can be seen as a genet­ic nose – a new way of smelling the bio­sphere. It can­not detect everything per­fectly, but it opens up pos­sib­il­it­ies that, com­bined with oth­er approaches, can only improve our under­stand­ing of biodiversity.