< Back to Resources
< Back to Viruses and Viroids

Variety Naming in Cannabis

Systematic Variety Naming in Cannabis

Why there's a necessity for it? What some approaches to doing that might be?

Dr. John Brunstein from Segra International

Source: ICR2020 Cannabis Variety Definition V5 – webinar presentation

Link: https://youtu.be/uvlx0E5NZD0

Segra is a Canadian-based Cannabis tissue culture nursery business. We grow plants at industrial scale for LPs. We can also do preservation and regeneration services by tissue culture for people and I oversee a genetics lab which performs various tests from everything from pathogen detection to what I’m going to talk about today, which is, genetic identification of plants.

We do this both internally and as an external fee for service. There are other methods other than the way we do this to do this approach and I’m going to try to be as even-handed as possible and discuss some of them as well.

One other disclaimer before I move on is that the term "strain" technically should be used for microorganisms so I’m going to use the terms variety or cultivar throughout this talk except in cases where I’m directly quoting. Customers buy cannabis based on name, in both the old and newer versions of Leafly, a common website that people go to for looking up cannabis it ties variety names to crowd-sourced reactions, so there's a linkage between name and effect.

Scientific data is similarly paired to names genbank, last time I checked had nearly 90 000 cannabis entries with self-identified names.  We see names used in biomedical publications, as well so this is a prevalent thing to do.  Names have meanings in other organisms, we can tell one kind of dog from another. We can tell one kind of grape from another and the type of wine that comes out of that grape can be quite different.  Between a Petite Sirah and a Muscat, and people pay differently depending on what it is.

Well could we also do that in a green plant diocese where the female inflorescences contain some sort of terpenoid compounds that we want to use in an industrial scale? Such a plant could be hops. And the answer there is absolutely yes. Hops have distinct varieties, those varieties have distinct reproducible characteristics, and if you buy a nugget or a citra or a bravo hops, you know what its characteristics are going to be.

Cannabis has a lot of variety names as well, but most these are self-reported. One clone could be sold under more than one name. Or you could have different clones, sometimes more popular name get a lot of things could get sold as that.

Varieties name for a cross are frequently named by conjunction of the parental variety names. So two different breeders doing a similar cross could produce an offspring product with the same name but these are not identical. They're siblings not twins.

Accidental or willful mislabeling, either way sometime in the past, can of course propagate forward. So even materials now in robust track and trace systems, depending on where they've come from, may not always be reliably named. Back in 2017 we wanted to know how many names there were, so we did a review of online dried flour for sale in a couple of Canadian and American jurisdictions. We looked at a total of 2739 items and we found 1263 names, but nine names alone were already 10% of the total population.

At the other end of the scale, 845 names only occurred once. It's unlikely all of those were actually unique.

So the question is, do these variety names have a reliable meaning in cannabis? Well how can we tell?

DNA fingerprinting is an obvious approach. No, plants don't have fingers. But a fingerprint in this context is a reproducible biometric mark of identity. Like a real fingerprint, which doesn't tell you if the person has blue eyes or blonde hair, a DNA fingerprint may not necessarily disclose physical phenotypic parameters, and we'll talk a little bit more about that later.

The method we use in our laboratory is VNTR profiling.  This is based on a small number of highly size variable loci and it's the method used in a lot of human forensics because it's reliable it's robust. It has small data sets and it's easily analyzed for identity and relatedness. Luckily, suitable loci already existing cannabis and are well understood in the literature.

Cannabis VNTR - Multiple Loci
Cannabis VNTR - Single Locus

You don't have to be a specialist to understand what the data looks like.  The blue peaks here are measurements of a particular size of loci and three different cannabis samples, and this one has two at 90 and 101 base pairs. This one has 2 at 90 and 96 base pairs and this one is a homozygous or possibly hemozygous 90, 90, and even to the non-specialist, this doesn't look like this doesn't look like this. Those are different clones. We take in now multiplexes across multiple  loci. In our case we do 12 different loci, we generate these numbers in order, 24 numbers. This is something we call a VNTR vector.  Vectors are identical between identical clones and they're different between different materials.  The probability of identity or the chance that two any such strings of numbers are the same between two things that aren't related is vanishingly small, somewhere between 1 and 14 million,  and 1 in 10 to the 49th of the two extremes. Theoretically for the sample we do and in reality for an average specimen if you look at the math, it's about one and a hundred million chance of a random match. This sort of data can be get phylogenetic trees as one way of visualizing data. And of course, tree topologies are models the actual shape of the tree is not always correct but absolute distances are true you can easily tell things that are clonal, from things that are clones to near neighbors, and you can tell things that are near genetically versus fargin.   So let's go back to our question.

Are cannabis varieties named reliably?

Well sometimes yes. So these two Shishkaberries are clonal, those two Shishkaberries are clonal, and these two and those two really on the scale of the entire phylogenetic tree are very close to each other. These two things are both called Cheese of some variety, they're not identical but they're really close.  These two things called Green Crack, down here, were in fact clonal but often you find examples of one name different genotype and, I won't go into details but in this case chemotype as well, all of the things shown there on the tree the, green star is three examples closely clustered something called Girl Scout Cookies and the red stars are each examples of something else called that they're not all Girl Scout Cookies.  Blue Dream,  Blue Dream was the single most common name in our sample, it was 2.1 percent of the entire population we saw for sale, three different things here called Blue Dream not close to each other at all.  I don't know if any of them are the right Blue Dream, but they're certainly not all the right Blue Dream.  You can see the inverse of that in this case.  We have one particular clone, here on this big isagenic line, we've seen at this point a total of 22 names applied to that clone.

Cannabis Phylogenetic Tree
Cannabis Phylogenetic Tree

If we do replicate sampling of legal stream material, we took 10 dried flower products purchased in the BC government chain each tied to a producer and a product name. We fingerprinted them, we waited two to three months and we repurchased the same product, same LP, again did a fingerprint and we compared. Well five out of the ten samples turned out to be the same lot number when we got them again, so mostly we learned that those products didn't sell extremely well, but all five of those scored isogenic to prior sample.  The other five however, were new lot numbers. Three of those five were isogenic, one wasn't isogenic to its prior sample but was close, but one out of the five was nowhere even close to its previous material.

What if we now do replicate sampling on single vials? Well we tried that too. BC government legal supply chain, one gram samples of dried flowers, sold as a single product variety name. We took multiple samples from distinct buds in the container, we VNTR them. What did we find? Well, out of 12 samples tested, most of the time one vial has one genotype but one vial detected with three different genotypes.  One vile detector with four different genotypes and within vials these genotypes were not all near neighbors, they were quite separate genetically. Notably both of these vials turned out to be from a single licensed producer. I'll let you draw your own conclusions from what that might mean.

What are the costs of inaccurate naming?

Well on a recreational side it impacts, branding and customer expectations, if you don't get the same thing every time when you buy the same label. On the medical side it's maybe more serious because the variety in the dose titration is very personal to get a good therapeutic window. One can imagine seller liability if a person is taking cannabis for let's say chronic pain relief, doesn't normally get drowsiness, get something with the name they expect, takes it, gets behind the wheel of a car and it's the wrong thing they do get drowsy they get into an accident, imagine what could happen there.

On the research side medical research and breeding programs are both negatively impacted if we can't reliably share data between different places and different studies.  All of these suffer unless varity names are applied consistently.

Really why should dogs or hops or grapes or roses have reliable breed of varietal names but cannabis. So what makes a variety? Well, it's a commonality of phenotypic traits of importance. Both of these in this picture we would call dalmatians because a breeder's association has decided that brown versus black spots is not a significant variation but that's a personal choice. And this sort of choice has to be made by type or reference examples by some body of authority. There needs to be an allowance for a certain range of phenotypic diversity, brown versus black here, arising from nurture or in this case a limited genetic diversity under a single appellation. Basically genetics provides a potential for the phenotype nurture can fine-tune the expressed phenotype but people have to determine which phenotypic variations are significant enough to warrant a new name.  

How do we measure genetic diversity?

Well there's multiple methods available, they're all molecular biology based, RAPD and ISSR are very low-tech, VNTR, the method we use here. There's also things like single nucleotide polymorphism or snap panels and genotyping by sequencing. Each of these has pros and cons, in particular I would suggest though, that SNP panels and genotyping by sequencing have significant capacity to disclose linked phenotypic traits.  In other words unlike a fingerprint in our analogy, SNP or GBS would tell you that the person had blue eyes in blonde hair, or in this case they might disclose detailed aspects of the phenotype, the chemotype, the disease resistance in a plant. A breeder might not want to be disclosing that, they just want a way to get a handle on identity and that's where we think VNTR kind of hits the sweet spot.  Multiple methods can of course coexist if they share an exemplar library.

How do we interpret genetic diversity?

Well phylogenetic trees are one way but, what about principal component analysis? PCA is hard to visualize meaningfully in two or three dimensions because it's multi-dimensional but it does allow for computation of point-to-point distances between any two genetic points, and it allows for relatively easy computation of self-asserting clusters or groups with centroid. So a cluster of things that maybe you want to apply a name to.  Alternatively, you can see this with a range of genetically distinct samples that you have said, I know what the name of this this and this are and now you have specific points and you accept things around that point.

Well accepted by who? Again goes back to this need for a body. The other side of course is chemotype.  Chemotype can also be done by multivariate chemotypic analysis. For cannabis you likely want to use both GC-MS and HPLC because of the range of volatility and polarity of the molecules you're looking at. But the results are equally amenable to PCA, and again you get this ability to have self-associating clusters and centroids. Or you can have a body provide reference materials and make these centers and decide what variations are acceptable. In fact there's a group out of Holland that's done this kind of analysis, Haze Camps group, one of their papers, they claim that chemotypically there's only really eight cannabis hemotypes.

Well moving forward, What should we do?

I think the question here is, what constitutes a suitable body? Probably I would suggest a group of recognized authorities, some long-term breeders, seed banks, members likely spread internationally.  

What do they need to define?

Well type specimens for extant popular variety names. Some sort of methods or metrics for both genetic fingerprint and chemotypic analysis, and then finally inclusion or exclusion criteria based on both genotypic and chemotypic similarity to a reference centroid for a variety. If you're too different genetically or too different chemically you're not the same variety.

Well for new varieties this is probably not too hard to handle because there could be some way to submit a sample for genetic fingerprinting and chemotypic signature to the body and there could be issuance of some sort of recognition of a variety name. Some sort of certificate. Extent varieties well some have traceable original clones and you can go back and do this. If not, maybe a majority rules from members of this body or panel.

What about other options? I think this is a topic for discussion too long to get into in this talk but hopefully this stirs some thought about that.

What are the challenges and benefits of doing this?

Well the challenges primarily, I think, are that some producers might need to rename a current product to harmonize with the rules. But the benefits for breeders includes protection of novel varieties. For consumers, medical and recreational improved uniformity of product.  For researchers, improved capacity to find correlation between variety and effects. For producers, an ability to market based on a provable uniqueness, " hey our product is different than everyone else and here's the proof".

Has this been done before?

I won't go into the whole story but in human immunology it's been done.  People all over the world were working on molecules they were naming them whatever it was in their own lab, and using antibodies to identify them.  Eventually they got together, pooled their antibodies and then found the ones that all bound the same thing and called those Cluster Designation one, two, three, four, now known as CD4, CD8, CD22.  You can pick up a newspaper and see these terms now.  It is possible to get everybody to rename in the same fashion.

Final thoughts

Is there an alternative?

Yes we can all choose to simultaneously believe multiple mutually contradictory things, which are cannabis variety names as they currently exist. If we do that I’m going to suggest we all need to get an Electric Monk. For those of you who don't know what that is, this is a handy device suggested by Douglas Adams and to quote partially here,  "Electric Monks believe things for you, thus saving you what is becoming an increasingly onerous task.

And with that I'd like to give my acknowledgements obviously the people at Segra a special thanks to Jerry and Ying, May, Manmeet and Paramjeet in my laboratory, and you the audience for your attention.

Viruses and Viroids

Tomato Bushy Stunt Virus in Cannabis

Read Article

Viruses and Viroids

Tobacco Streak Virus in Cannabis

Read Article

Viruses and Viroids

Lettuce Chlorosis Virus in Cannabis

Read Article

Viruses and Viroids

Hop Stunt Viroid in Cannabis

Read Article

Viruses and Viroids

Beet Curly Top Virus in Cannabis

Read Article