Humans and chimpanzees are often described as cousins, but that analogy doesn’t quite capture the true distance in the relationship. Cousins, after all, can be kissin’ — and humans and chimpanzees aren’t into that. And, even if we were, we couldn’t make a baby. In other words, we’re different species. And yet, most of our DNA is the same. The human genome and the chimpanzee genome really only differ from each other by about 1.2 percent. The lesson: Even teeny little differences in genetic variability can matter a lot in outcome.
Which is why any discussion about the genetic variation in cannabis — the plant more commonly known as the sedate “hemp” or the sedating “marijuana” — can quickly become mind-blowing. “We have found a lot of variation in the plant,” said Daniela Vergara, a postdoctoral evolutionary biologist at the University of Colorado at Boulder. “We have found as much genetic variation as there is between different species, between chimpanzees and us.” Yet, somehow, despite that diversity, cannabis is still one species. All those plants, even the ones as different from each other as humanity is from chimpanity, can still make babies.
In a word: Whoa.
As of November’s election, eight U.S. states have legalized both recreational and medical uses of cannabis, although it’s still illegal at the federal level. More states have now completely legalized the plant than continue to completely ban it, according to the National Cannabis Industry Association. As cannabis creeps back from nearly a century of criminality, scientists at marijuana biotech startups are starting to analyze its DNA, hoping to understand better how the plant works and help bring some scientific legitimacy and consumer protections to the fledgling industry growing alongside it. What they’re finding is that the plant’s astounding amount of genetic diversity is complicating cannabis use, threatening not only consumer confidence and safety, but also the accuracy of scientific research.
Genetic diversity is what first made cannabis popular with humans — making it useful in a variety of ways, from drug to textile. And the plant’s genetic diversity has increased over time, as humans bred it and carried it to new ecological niches. There are samples of hemp textiles from China that are 6,000 years old. There’s evidence that ancient Egyptians used it in insecticides and pain medication for births. By 2,700 years ago, the ancient Chinese had plants that contained high levels of the psychoactive compound THC, suggesting that they could get high and were breeding cannabis for that purpose. Although it’s likely that cannabis originated from plants that grew in central or eastern Asia, it now grows outdoors on every continent except Antarctica. It grows near the Arctic Circle. It grows in Mexico. That’s unusual for plants.
Cannabis is also unusual because of its criminal past (and present) — a history that cut it off from modern agricultural science during a crucial time for the understanding and documentation of genetic diversity in plants, and for putting that diversity to work for consumers through focused, coordinated breeding programs. The nearly 100 years that cannabis has been criminalized overlapped with the Green Revolution — a worldwide period of plant research that resulted in the creation of seeds as reliable, branded products that produced the same traits every time they grew. Cannabis, confined to basements and back alleys while other plants entered the laboratory, missed out on this, said Mowgli Holmes, chief scientific officer of Phylos Bioscience, a cannabis biotech company.
So while other domesticated plants have genetic lineages that are carefully documented, cannabis … doesn’t. Companies like Phylos are just starting to do that research. Basically, cannabis the plant might be legal now in some places, but cannabis the industry is still playing scientific catch-up.
One result of this is that consumers never really know what they’re buying, even in states where cannabis is legal. Cannabis is sold under names — Sour Diesel, Girl Scout Cookies — that imply standardized varieties, the same way you can buy Xtra Sweet corn or a Bushel Boy tomato. But the science doesn’t really work that way, Holmes, Vergara and other scientists said.
Holmes’s company has an online data visualization showing genetic relationships between samples of cannabis DNA that have been submitted to it by people from all over the world. About 20 percent to 30 percent of the time, he told me, the samples they test won’t be even remotely related to other samples sold under the same “variety” name. That’s similar to results that a Canadian research team got in 2015, when it found that 35 percent of its samples were more genetically similar to samples with a different name than to samples with the same name. Holmes said that these distinctions can matter to recreational consumers who had a bad experience on a certain variety and want to avoid it in the future.
That also applies to people who are picking their weed based on the broad “family” categories of sativa and indica. The former is often said to produce highs that make you feel energetic and creative, while the latter is supposedly the stuff that makes you fall asleep on the couch after 15 minutes. But “it’s all just a big myth,” Holmes said. And other scientists agreed. The 2015 Canadian study, for instance, found that even though there is some evidence that cannabis could be organized into two families with distinct traits (and even that is the subject of scientific debate), those “sativa” and “indica” designations aren’t really meaningful to the consumer market, because of both cross-breeding that muddies the distinction and simple mistaken identity. In one case, a variety called Jamaican Lambs Bread, which was supposed to be a sativa, was nearly genetically identical to a supposed indica variety from Afghanistan. Both the names of cannabis varieties and the indica/sativa division should be thought of as fungible marketing, not accurate biology.
In states where cannabis has been legalized, there are a variety of regulations governing what must be on the drug’s label and what can’t be there — but other than prohibitions against untruthful statements (which only work if the seller knows they’re untruthful), these rules don’t require sellers to ensure that their Sour Diesel is actually Sour Diesel. Largely, Holmes told me, this is because nobody has the authority (or, for that matter, the right) to say what the official Sour Diesel is. With no historical or legal documentation, there’s no way to know which of two genetically different samples sharing a name is the “real” one.
These issues have big implications for the medical cannabis industry. That business relies on being able to sell named varieties that are high in CBD — a naturally occurring compound in cannabis that doesn’t make you high but might be able to do things like reduce the symptoms of epilepsy.
In 2014, Verda Bio, a different cannabis biotech startup, tested 22 samples of Harlequin, a variety that is supposed to be high in CBD. Twenty-three percent of those samples contained almost no CBD — but were high in THC. In other words, the name “Harlequin” doesn’t tell you what you’re getting, and a child using this variety to prevent seizures could easily end up stoned instead. Some states, including California, require sellers to put information about the levels of pharmacologically active substances on the label. But not all states with legal marijuana have a mandate like that.
The scientists studying how cannabis affects seizures are also running into genetic diversity problems. In their case, though, the problem is that the cannabis they’re allowed to study doesn’t have the diversity seen in the real world.
The only cannabis that is federally legal for scientists to study comes from a supply maintained by the National Institute on Drug Abuse. In a forthcoming research paper, Vergara and her team compared samples of cannabis from the institute’s supply with samples from stores in Denver, Seattle and the California cities of Oakland and Sacramento. Compared to the commercial stuff, the institute’s samples had much less diversity, in terms of the quantity of THC and CBD. That’s good for research replication — unlike shoppers, scientists can take two samples of cannabis and trust they’ll be fairly similar to each other.
But it’s bad for figuring out what cannabis can really do for human health. Vergara’s team found that the institute’s samples have, on average, significantly lower levels of THC and CBD than the samples bought in stores. That matters, because we don’t yet know how the extreme genetic diversity of cannabis affects the human body. For instance, consider the question of whether CBD really does reduce epilepsy symptoms. Most of the data supporting this connection is anecdotal, and the results of clinical trials have been mixed. But Vergara found that the percentage of CBD in the National Institute on Drug Abuse samples was lower than levels in almost anything an epilepsy patient could buy in the real world. If scientists can study only one extreme on a wide continuum of possible doses, there’s no way to know whether a study that fails to show cannabis providing relief from seizures means the drug doesn’t work or whether it just means the scientists’ samples aren’t diverse enough. What’s more, there are a whole host of less-well-known chemical compounds in cannabis that probably affect both the effectiveness of medicinal use and the effects of recreational use. Positive test results — and, thus, scientifically proven claims of medical efficacy — could depend on how those compounds interact with the more-well-known ones.
Companies like Phylos and Verda Bio are doing work they hope will lead to stable cannabis hybrids and a certification system that would allow sellers to easily understand and transparently document genetic relationships between their plants and others of the same name — changes they say would increase consumer protection by helping people know whether they are getting the genetic traits they want, or need. Holmes even imagines a future where cannabis genetics will be so well-researched that breeders will be able to link genes to both the plant’s own physical traits and the types of experiences it creates for people who use it. Wouldn’t it be nice, he suggested, to have a variety of cannabis that reliably made cancer patients hungry without leaving them spaced out … or a variety that produced a mild high equivalent to a single glass of after-dinner wine?
But Vergara’s data demonstrates how state-based legal cannabis — its use and how much trust consumers can place in their suppliers — is still heavily influenced by the plant’s clandestine history and its continuing illegal status at the federal level. What good is legal medical marijuana if consumers have no way to verify that they’re getting what they paid for? What good is a well-labeled, verified, high-CBD variety if scientists can’t study it to explore its health-care claims? There’s a weedy tangle here, combining genetics, the needs and wants of consumers, and the law. To get to Holmes’s imagined future, the cannabis industry and its advocates will have to find a path through.