Genetics to replace morphological examination?

Discovery of ‘cryptic species’ shows Earth is even more biologically diverse | Wildlife | The Guardian
I came across this yesterday, and while I am sceptical of a ‘genes only’ taxonomy, I would be interested to hear what others think.


Fascinating article. Thank you for sharing.

Overall, I think morphology is still an incredibly important part of identifying wildlife and organizing taxonomy, especially among the organisms most often encountered and photographed by people (e.g. vertebrates, plants, fungi), and I do not see that ever going away or being mitigated. Taxonomic studies based on genetic information are sometimes spear-headed by preexisting observations on divergent morphology, so I see them as supportive of one another. Even in the absence of morphologic info, usually a split between taxa based on genetic information often are at least also geographically isolated from one another or separated by other environmental conditions (e.g. trophic level, altitude, diurnal/nocturnal, etc.). Organisms can be identified through those means as well. Even among invertebrates, which often have more cryptic diagnostic traits, are still subject to taxonomic revisions based on morphology (e.g. Huang, 2017, who split up the traditionally monotypic Hercules beetle into several species based on the shape of their horns and geographic distribution).

From a practical point-of-view, for naturalists, I do not think genetics will ever replace morphological examination. Your average naturalist - who may not be a biologist with access to genetic-sequencing technology - is not going to benefit from that information when identifying wildlife. In a world where genetic examination becomes preferred over morphology, the worse case scenario is that naturalism becomes increasingly insular due to overreliance on genetics, making resources like iNaturalist less accessible and welcoming to everyday folks. However, I do not see that ever actually happening.

Genetics is a helpful tool for contextualizing and organizing the natural world and the relationships between them. I do not realistically think it can be the only defining factor.


It seems that the general consensus is that each individuals DNA is unique.

With the possible exception of immediate family members, “shared” DNA does not absolutely prove a connection.

What percent “shared DNA” is necessary to determine if it is a different species.

It is my opinion that DNA is one of many criteria to be used in distinguishing different species.


Genetic, taxonomic, biogeographic, and behavioral analysis complement each other. Each is incomplete in and of itself, but taken together they paint a more complete picture. Genetic analysis is unlikely to replace taxonomic (or any other sort) of analysis, but it provides an important window into the differences between populations and species and is increasingly important due to its quantitative nature.

One of the big problems is, of course, interpreting the results of your analysis and what that means in terms of species delineations. Right now, and for a while, the “splitter” faction of species determination has been highly influential. I suspect that the pendulum will swing back a bit in the future once the novelty of easy DNA analysis wears off and we understand what’s actually important in differences in DNA better. Some of the species marked out at present based on genetic analysis as disparate species will likely be recombined again, given subspecies designations instead.

Just as in fashion there are trends in science; mindsets and accepted paradigms change and they swing back and forth when new ways of analyzing or conceptualizing things emerge.


Yes, and now with Transcriptomics, Proteomics, and Metabolomics, there is even more data to analyze! This means people also need to become more computer savvy to analyze the data. I will admit that I am behind in the computer savvy department.

1 Like

iNat’'s Computer Vision performs a very similar role to DNA Barcoding. For non-specialists, it’s a Black-box. You don’t need to know anything about how it works: you just feed it some data, and it gives you an answer. If DNA Barcoding technology ever reached the point where it could be packaged as a hand-held device that gave results in a few minutes (or even hours), I think it would rapidly replace most of the traditional methods. Whether this would be a good thing or not is somewhat moot. Practicality always beats purity in the long run.


To the extent that it means anything taxonomically, morphology is expression of genetics. Morphology that is not genetically based is not relevant to species identification. In a nontrivial sense, then, all species ID is about genetics. The so-called barcode is just another instance.

There are undoubtedly many questions to be asked of the barcode data, not least the question of how barcode species are related to other species defined using other concepts. Ultimately, this is just another species concept in a long and growing list. It may actually turn out to be more biologically meaningful than many of the others. In the meantime, it is a useful tool for understanding diversity.


No, not yet. I can tell a bit about the world of fungi. There was (still is) a great enthusiasm to describe species without having a specimen = from sequences. There are pros, because many species of fungi have no (known) fruitbodies of any kind and cannot be grown on media. So the only way to detect and name them is their DNA. But… there is obvious caution starting to get into the latest taxonomical papers, as it became obvious that quite a number of species previously segregated mainly on the basis of DNA sequences without paying sufficient attention to morphology, ecology and geography are indeed very flawed. On the other hand, at least several lumpings or splittings (mainly the latter) made solely on the basis of morphology were resolved thanks to molecular studies. Then again, several groups got totally stuck in limbo when it appeared that neither morphology, nor DNA studies (at the present level of knowledge) can solve species status. Ant finally metabolomics. Another great expectation, but just a year ago I heard first cautioning presentation at a conference. Metabolites, at least in fungi are not defined only by species genetics, but also by the environment and interactions with surrounding species. So, the 20 year old dream of my colleague of a little machine with which you touch lichen and it beeps “Ramalina pollinaria!” is still far ahead. So, naturalists –do not worry, your data is still needed. And will be – for a considerable time.


These articles are perennial (note the example from 2004, which is one I recall seeing before). See‘revolutionary’_methods_to_accelerate_the_description_of_species for a sharp critique of a recent (2019) attempt to do this for a genus of braconids.

I think in practice, even in groups which are largely morphologically delineated, we rely in practice on other cues (behavior, foodstuff, habitat, location…) to a certain degree in practical identification. Engineering is bringing a “DNA tricorder” closer to reality, but I still tend to think that the main role of sequencing will be to help us identify which non-morphological characteristics are signals of genetic differentiation rather than a standalone criterion for species distinction, at least in sexual eukaryotes.


So far, the responses echo my own beliefs - that what organisms ARE is more than the sum it it’s genes. I was first exposed to this in the Massey Lecture Series/book by Richard Lewontin (Biology as Ideology: the Doctrine of DNA). Although genetics have a lot to do with it, what an organism is is more than genes. People have mentioned behaviour, feeding and other behaviours as being part of taxonomy. I completely agree with this. Life forms are malleable, and to reduce them all to genetic information strikes me as being absurd. I have noticed this announcement on BOLD, for moths - " * Barcode of Life Data System (BOLD) - Caution: Some specimens shown may not be sequenced. DNA barcode provides evidence of relatedness not proof of identification. I know that Euxoa ochrogaster has many morphs that look radically different, but they can all interbreed.


Morphology, which is part of an organism’s phenotype, is what taxonomists used for centuries before allozyme and genetic techniques came along. But as we know, it can be deceptive when dealing with organisms that are polytypic or lack discernible or measurable physical characteristics that allow grouping/separation, or in cases of convergent evolution.

In an ideal world, where we had full understanding of every organism’s genome and the evolutionary history behind that genome, we wouldn’t need morphology – we could map out the placement of every lineage in the tree of life. Morphological differences among species would simply be a reflection of our taxonomy and not a driver of it. But, since nature is messy, it still would come down to some taxonomic decision-making to say what’s a species, genus, family, etc. Assuming Linnaean taxonomy still has a place.

We’e not there yet, of course – maybe we’ll never be – so I think morphological examination still has a place. But I can see the day when it will be looked on as rather quaint. I know some phylogeneticists who already think that way.


Two issues here. First, can genetics replace morphology for deciding whether two populations are one species or two? Answer: Sometimes. After all, different species no longer interbreed; they no longer exchange genes. Genetic analyses may be able to determine whether that’s true or not. However, there is no amount of genetic differentiation that indicates two populations are one species vs. two. Sometimes, a small genetic difference makes two populations unable to interbreed to produce fertile offspring. Other times, one species is highly genetically variable.

Sometimes genetic incompatibility results from translocations (rearrangements of gene order on chromosomes) or polyploidy (variation in number of copies of each chromosome). These can cause virtually instantaneous speciation, with, at first, no or few detectable difference in DNA sequences. And sometimes even polyploidy doesn’t prevent (though it does reduce) gene flow between, for example, diploids and tetraploids.

Sometimes there is clearly small morphological and genetic differentiation between two populations and it clearly results from geographic separation. Are the two populations one species or two? Would the organisms interbreed if they met one another? That’s a judgement call, and it can not be resolved with more data, morphological or genetic. Also a judgement call: where do we draw the line between asexual species? How do we handle populations that seem to be in the process of becoming two species but haven’t completed that?

Genetic analysis provides data for deciding what are species. Important data. But not a simple replacement for looking at morphology, behavior, etc.


Taxonomy is a human construct, merely a way for us to describe the differences in populations of organisms, that kind of thing. It began with differentiating by morphology, but we have other species concepts as well - biological, ecological, and phylogenetic are the other main ones. Which one you go with depends entirely on your goal. That said, phylogenetic is the most precise by far, and gets better the more data you have - a barcode is good, reduced representation sequencing is better, and whole-genome sequencing is the best (not taking into account how hard it is to actually interpret said data). In a similar way, the more morphological characteristics you consider, the better your ID.

Phylogenetics will never replace morphology, but morphology should not come before phylogeny, when it is available. Phylogenetic trees are a lot more useful when you map morphological traits on to them. Describing a new species (fungi, especially) on morphology without sequencing it first is real silly given how cheap and easy it is these days.

@bazwal Handheld quick sequencing device you say? It exists, and it’s pretty cool!


Second issue about genetic data: Once we’ve decided on the species boundaries (see above), can we rely on genetic data (including but not limited to DNA barcoding) for species identification? Can / will genetic analysis replace morphology, behavior, etc., for identification?

A note: DNA barcoding as I learned about it refers to using one gene or a few genes for identification. Could theoretically be simple. But which gene(s) work for a given group? For example, a gene that often provides species level identification of plants misses most of the species boundaries within the genus Carex. Probably other genes will work, and likely whole-genome sequencing would work, but it’s not going to be as simple as analyzing one snippet of DNA and getting species ID’s for all plants.

You probably know that a skilled birdwatcher, botanist, etc., familiar with the area can go in the field and (depending on the season) reliably identify most of the birds or plants or whatever by sight/sound and do so almost as quickly as he sees or hears the animals. (We can also ID many photos fast, too.) That’s great, it’s relatively cheap, and it doesn’t require handling the animal or crossing the creek to get a piece of that plant. The value of this skill will never go away.

There are problems with ID’s by morphology, behavior, etc., though. What about animals that are usually hidden? What about plants that aren’t out at the time you have to be in the field? What about molds and other organisms that nobody knows much about? What about cryptic species? In these cases, that DNA barcode reader could be a wonderful tool.

Another issue is the time it takes to master the identification. Becoming good at DNA analysis takes time. Becoming good at insect identification takes time. Becoming good at grass identification takes time. Learning to dissect insect genitalia and know what you’re looking at takes time. Most of these skills are highly specific to the group of organism and the place. DNA analysis can be generalized. IF IF IF IF IF somebody has worked out the taxonomy and found handy genetic traits to use to distinguish the species, a gene jock who can barely tell a heron from a sparrow can do the identification (assuming DNA samples are available).

So, I think genetic analysis is a useful tool and will only become more useful as the equipment and processes get cheaper and faster. But it won’t simply replace identification based on morphology.


It seems like genetic analysis probably poses the most threat to the more bottle-necked sorts of analysis, e.g., “The only way to definitely identify this braconid is to mail it to a septuagenarian in Prague with an 18-month backlog.” (Think of how, in clinical microbiology, PCR and immunological diagnosis have done away with a lot of culture-based testing.) But even with cheap, fast, easy sequencing, you don’t want to throw, say, every lichen you encounter into the machine and have 50% of them come back as Flavoparmelia caperata. I think having a human on the front end as a “pre-filter” for what you decide to sequence is always going to be useful.


Sounds suspiciously similar to AI suggestions :-)


That isn’t a black box device, though, because it requires specialist knowledge in order to provide the right input and then interpret the output. What I had in mind was a device that takes arbitrary organic material as input, and gives the nearest taxon as output (perhaps with some kind of confidence level). It’s really irrelevant how such a device would actually work. It might use DNA barcoding. But it could also conceivably use laser scanning to produce a 3D model for morphological analysis. iNat’s Computer Vision is (with all due respect) a “lite” version of such technology. It takes input via a camera phone, and an app outputs a suggested taxon.


All living things literally reduce themselves to genetic information every time they reproduce/replicate. Absurd maybe - but that’s life.


I just want to add that if even if you use a specific gene, there is variation in the length of the recovered DNA/RNA; depending on the length of recovered gene/gene product, the phylogeny of the same organisms may or may not change.

I was surprised when I tried to compare an organism from all taxonomic ranks in NCBI BLAST. Instead of having the phylogeny that I expected, where every organism in higher taxonomic ranks are more closely related to each other than those of lower taxonomic ranks, the phylogeny was scrambled. Now, this may have to do with the huge variation of relatedness among the organisms, which is beyond the scope of any BLAST analysis. It is possible that other phylogenetic programs, like CLUSTAL, will give the expected phylogeny. However, this does show a limit to phylogenetic analysis.


I won’t argue with the reproduction aspect - all life is based on genetics. However, to reduce all life to genes ignores the influence of climate, environment etc. on the expression of genes. I’m not a geneticist, but the variation in behaviours expressed by life forms in different circumstances seems to indicate a more complicated interaction between an organism, the environment and genetics. You and I have basically the same genetic makeup, but I would be willing to bet that we are different people who interact with the world differently.