Taxonomy: Part 1. What is it?

I have been asked to write a few words about biological taxonomy – simple, heh! If only! Firstly, as naturalists and biologists, we probably think mostly in terms of genus and species. But – taxonomic classification is just a system whereby things are organized into groups or types. The Dewey System of numbering books in a library, or considering traffic as travelling on freeways, highways, roads, lanes, or tracks is also a taxonomy. Using the Dewey system as an example, do we put a book called “Trees in the Landscaped Garden” into “trees”, “landscaping” or “gardens”? These are all legitimate Dewey categories. Using the road example, the difficulties become more obvious; how narrow does a road have to be before it becomes a lane? How narrow or rough does a lane have to be before it becomes a track? The same problems apply (very much so) to biology. Plants are green and stand still – right? Animals are not green and move around – right again? So, what do we do with animals that move around and are green because they have chlorophyll?

By Barry Muir

The big difficulty with biology, as with roadways, is that the concept of species and all the other “higher” classifications, phylum, family, etc., is that they are all artificial, human-made concepts and may bear only a limited relationship to real life. That is why species names change; Fred looks at some characters and decides it is species A, then Jennifer looks at a different set of characters and calls it species B, then Jock looks at it and decides it is really just a variant of species C, so lumps B and C together. Taxonomists have been known to come to blows over such distinctions. Some like to create new species (the splitters) and others like to group them together (the joiners).

The bottom line is that these classifications are primarily a human convenience. When studying things, we need to be able to distinguish between a horse and a llama, so we give them different names. Horses and llamas are OK, but what happens when we deal with a fungus? These are tricky because there are really very few physical characters, gross or microscopic, we can use to distinguish one from the other and even those that are distinguishable may be highly variable within a species.

Mycologists have traditionally used morphology (physical characteristics), such as spore-producing structures, as a means of identifying fungal species. Structure is very useful for classification of fungi at the order or family level but may not always perform well for genus or species. Physical characters can be misleading due to hybridization, complex speciation, or even convergent evolution, where unrelated species evolve very similar characteristics. Consequently, we turn to DNA sequencing – that’ll make everything clear – right?

DNA may be used to determine the sequence of individual genes, larger genetic regions (i.e., clusters of genes, full chromosomes, or entire genomes) of any organism. DNA sequencing has become a key technology in biology and sciences such as medicine, forensics, and anthropology. It has become a household word since Covid-19 appeared. Have a look at Wikipedia – DNA Sequencing, for more information. Sequencing has even resulted in a “new” taxon, called the Operational Taxonomic Unit or OTU – sort of like a species but based on DNA.

New fungal phyla, classes, orders and families have been established in the last decade by means of DNA studies. Additionally, the fungal species on the earth has now been estimated to be 12 (11.7–13.2) million compared to previous estimates of 2.2–3.8 million species (Wu, 2019). However, DNA sequences cannot be solely used to assign scientific names to fungal species until physical voucher specimens, or even illustrations that can act as the type specimen, are available. A type specimen is a single specimen (or equivalent) which sets the benchmark for that species and against which all subsequent specimens are compared.

To oversimplify, two methods of DNA analysis are mainly used in mycology; what is known as DNA barcoding and another using one or multiple genes in sequence and then using special tools to estimate their relationships. See Raja et al. (2017) for a very detailed explanation of how this works.

In DNA barcoding, the user compares an unknown DNA sequence against a published database and identifies species based on similarity. Some barcodes are not useful for determining family and order levels. The premise of DNA barcoding is that variation between species should be greater than the variation within species. This is not always so – making for occasional unreliable identifications.

Another approach uses a special region of DNA called the Internal Transcribed Spacer (ITS). This is the “official” DNA barcode because it has been found to be among the genetic markers with highest probability of correct identifications for a very broad group of fungi. However, even though the ITS method performs well as a suitable fungal barcoding marker, it has been subject to debate. The ITS region, for example, does not work well in some highly diverse mould genera, such as Aspergillus, Cladosporium, Fusarium and Penicillium, or with Trichoderma.

The next question is – how similar is similar? Orang Utans and humans have only 3 % of their DNA different yet are clearly separate species. No “cutoff” value of similarity has been universally accepted. In the past, mycologists have used an arbitrary cut-off value ranging from 3% to 5 % for ITS sequences to indicate same-species among fungi (i.e., up to 5% variability in the sequence) for assigning a species name (Raja et al., 2017).

Another complication is that fungal DNA sequences are not always recorded in official databases and their accession numbers included in manuscripts. This makes it very difficult to link published taxonomy to genetic sequences. Verification of identifications is crucial but often neglected (Lucking et al. 2020). Approximately 27 % of database fungal ITS sequences have been found to be submitted to databases with insufficient taxonomic identification, and about 20 % are believed to be erroneous. Some taxonomic names are also not up to date due to the rapidly changing nature of traditional (morphological) fungal taxonomy. Most fungal species described using the old methods (about 70 %) thus far have not been DNA sequenced and many sequences are unnamed or only partially named. This results in many unpublished and unreliable DNA sequences (Raja et al., 2017, Wu et al. 2019).

To put this technical jargon more simply, if we have DNA from an unknown organism that very closely matches the DNA from a thing we call a horse, then it must be a horse. The problem arises when it is just a bit different! If it is 98 % the same as a horse we might call it a horse, but if it is only 97 % the same is it still a horse, or is it something else? In reality, and somebody is sure to take me to task on this statement – classification by DNA is, in a way, just as arbitrary. It is generally agreed by all mycologists that, whenever possible, identification of fungi should be made using a combination of structural, ecological and DNA information. Even forensic DNA analysis, which is extremely carefully controlled, is only considered 95% accurate by some specialists. A great deal of DNA work in universities is also done by inexperienced students, perhaps reducing reliability.

Stay tuned for Taxonomy: Part 2. The Tree of Life next month!


Lucking, R. et al. (2020). Unambiguous identification of fungi: where do we stand and how accurate and precise is fungal DNA barcoding? IMA Fungus 11:14 32pp.

Raja, HA, Miller, AN, Pearce, CJ. & Oberlies, NH. (2017). Fungal identification using molecular tools: a primer for the Natural Products Research community. J Nat. Prod. 80:756-770. doi: 10.1021/acs.jnatprod.6b01085. ( np6b01085.pdf)

Wu, B, Hussain, M, Zhanga, W, Stadlerb, M, Liua, K. & Xiang, M. (2019). Current insights into fungal species diversity and perspective on naming the environmental DNA sequences of fungi. Mycology 10(3):127–140.