Why the THC:CBD ratio is in the genes

Every seed catalogue sorts its cultivars the same way: THC-rich, CBD-rich, or "balanced" 1:1. And on forums, the same questions keep coming back. Can I make my plant produce more CBD? Will extra light, a dose of UV or some deliberate stress push it one way or the other? Does harvesting late change its character? Behind these questions is a reasonable intuition. Growers know that the environment shapes almost everything about a plant: height, leaf size, yield, aroma. So why not its chemistry?

The research gives a fairly clear answer, with some honest gaps. The amount of cannabinoids a plant makes responds strongly to how it is grown. The ratio between THC and CBD, the plant's chemotype, is set largely by its genes. This article explains why, looks at the studies behind that conclusion and what they actually measured, and considers what it means for someone choosing seeds for a 1×1 m tent. A legal note: in the UK, growing cannabis, even with a prescription, requires a Home Office licence; this article is educational.

One precursor, two enzymes

The branch point

Cannabinoids are made in the glandular trichomes, the tiny resin heads that are densest on the bracts of female flowers. Almost all of them start from a single molecule, cannabigerolic acid (CBGA). Ahsan et al. (2026), in a review of cannabinoid genetics, describe how it is assembled: olivetolic acid, built by a short chain of enzymes, is joined to geranyl pyrophosphate, a building block from the plant's terpene pathway. CBGA is the point at which the plant commits to making cannabinoids.

What happens next decides the chemotype. Two enzymes compete for the same CBGA. THCA synthase turns it into tetrahydrocannabinolic acid (THCA); CBDA synthase turns it into cannabidiolic acid (CBDA). Ahsan et al. describe them as "substrate-competing enzymes" whose relative expression determines the CBD:THC ratio of a cultivar. In the living plant, both products stay mostly in their acid forms. They lose a carboxyl group and become the familiar THC and CBD when heated, dried or aged.

A useful picture is a single pipe that splits into two taps. How much flows through the pipe, the total amount of cannabinoid, depends on many things, including the plant's size, health, light and stage of development. Which tap is open, and how wide, depends mainly on which versions of the two synthase genes the plant carries.

Working and broken copies

A gene can exist in working versions and in broken ones. A broken version (a pseudogene) may have a small typing error in its DNA, a missing letter that throws the rest of the code out of frame, or a larger defect. Ahsan et al. (2026) summarise the evidence that inactivating mutations of this kind "determine whether a given synthase allele is functional and thereby set the qualitative drug-, hemp-, or CBG-dominant chemotype". The synthase genes sit in large, repetitive stretches of DNA that differ in structure between drug-type and hemp-type plants, and each region contains one active enzyme alongside numerous inactive copies. The number of copies also varies between plants. The two synthase genes are tightly linked, so they tend to be inherited together. The review refers to this linked region as the B locus: in its words, structural and copy-number variation, the functional or pseudogenised state of the synthase alleles, and "linkage at the B locus" account for much of the difference between drug-type and hemp-type chemistry.

The authors are careful about what genetics does not explain. Total cannabinoid content can vary "by as much as 10-fold across cultivars", and part of that variation occurs between cultivars with near-identical working synthase genes. The review's own subject is the hypothesis that the packing of DNA (epigenetics) explains some of this residual variation in amount. Those ideas remain largely untested in whole cannabis plants. In other words, genes decide the type with some confidence, while the amount is a more complicated story.

Chemotypes I to V

Henry et al. (2020) use a classification by the ratio of CBD to THC:

Henry et al. note that a genetic basis for types I–III "can be determined by polymorphism at the CBDAS and THCAS genes on Chromosome 9". They describe type IV plants as carrying non-working versions at this locus, so that CBGA builds up because neither enzyme converts it, and type V plants as having lost function further upstream in the pathway.

A simple consequence follows. Since each plant carries two copies of each chromosome, a plant can carry two THC-type versions, two CBD-type versions, or one of each. In general genetics terms, plants with one of each tend to make both cannabinoids, which is the usual explanation for type II. Henry et al.'s marker data, described next, are consistent with this kind of simple inheritance at one stretch of chromosome 9.

What a genetic survey of 420 plants found

The study

Henry et al. (2020) genotyped 420 cannabis samples from licensed cultivators in Saskatchewan, Manitoba and British Columbia in Canada and in Nevada in the United States. They used a panel of 23 single-nucleotide markers: single-letter variations in DNA chosen because they are linked to traits such as cannabinoid and terpene production. Cannabinoid and terpene data were available for a subset of 118 resin-type samples. Several authors work for genetics and testing companies.

Markers that track the chemotype

Three markers in the THC synthase gene cluster on chromosome 9 strongly separated the major groups. Two of them, SW6 and VSSL_BtBD, "were fixed for one allele in all CBD expressing varieties (fibre and resin-types), while being fixed for other allele or heterozygote in all THC expressing varieties". In plain words: every CBD-producing plant in the survey carried two copies of one form of this marker, while every THC-producing plant carried either two copies of the other form or one of each. That is what you would expect if the chemotype is mainly inherited and linked to this stretch of chromosome 9. It also means a chemotype can be predicted from a scrap of leaf. Henry's team needed only a 2 mm² piece of leaf tissue per sample.

Henry et al. also report that the active THCA and CBDA synthase genes are thought to lie a short genetic distance apart on chromosome 9, in a linked region surrounded by mobile DNA elements, and that an inactive CBDA synthase gene sits close to the active THCA synthase gene. This matches the "linked unit" picture from Ahsan et al. (2026).

Five groups, and mislabelled plants

The markers sorted the samples into five genetic groups: European hemp, high-CBD "resin hemp", and three THC groups that roughly matched popular "indica" and "sativa" labels and differed in dominant terpenes. High-CBD and balanced drug-type plants were spread across several of the resin groups. The authors read this as evidence that high-CBD drug-type cannabis has been bred from several different backgrounds rather than one.

For seed buyers, two findings stand out:

• Names do not guarantee identity. Among 420 samples, the team found 361 distinct multi-marker genotypes. Plants sold under different names often had identical genotypes: "Gorilla Glue" and "Holy Grail", for example, or "Atomical Haze", "Tangerine Dream" and "SFVOG". With 23 markers, identical results strongly suggest the same clone under different names.
• Seed lines are variable. The average heterozygosity within seed lines, a measure of genetic variety inside one line, was 0.33, which the authors considered "much higher than what was to be expected in any other major stable commercial crops". A Canadian fibre and grain hemp cultivar was the most uniform line, at 0.09. Some drug cultivars from known seed banks were fairly stable, below 0.2. Others showed signs of mislabelling or widespread use of first-generation (F1) hybrids, which do not breed true.

The survey covers North American material from licensed producers. The underlying genetics should be the same elsewhere, but it says nothing about how mixed or mislabelled UK and European seed lines are.

What the environment can change

If genes set the type, what does the environment do? Three studies of whole flowering plants are useful here, and they point in the same direction: growing conditions change how much the plant makes, not the THC:CBD ratio.

More light: more flower, same chemistry

Rodriguez-Morrison et al. (2021) grew the high-CBD cultivar 'Stillwater' in an indoor commercial facility in Southern Ontario. Rooted cuttings in peat-based medium in 1-gallon pots sat in deep-water basins, about 11 plants per m². After two weeks of vegetative growth, plants flowered for 81 days under a 12-hour photoperiod, with LED light at canopy level ranging from 120 to 1,800 µmol/m²/s across the plants. CO₂ was ambient, at around 440 ppm during the day, and air temperature was about 25 °C.

The effect on yield was large. Dry flower yield rose linearly from 116 to 519 g/m², 4.5 times higher, from the lowest to the highest light level, and cannabinoid yield per m² rose by the same factor. But cannabinoid concentration did not change: "There were no LI treatment effects on cannabinoid potency." The flowers contained about 214 mg/g CBDA and 12.9 mg/g THCA, a ratio of roughly 17:1, with no difference between light levels. More light meant more flowers. It did not mean different flowers.

More light and UV in a THC cultivar: still no CBD

Light trials with high-CBD cultivars leave an obvious question: could a THC-dominant plant be nudged towards CBD? Llewellyn et al. (2022), working with a Canadian licensed producer, grew the high-THC cultivar 'Meridian' in enclosures in a commercial facility, also in Southern Ontario, at about 10 plants per m² in rockwool. Plants flowered for 45 days under 12-hour LED light at 600, 800 or 1,000 µmol/m²/s, or at 600 plus extra ultraviolet: UVA for the whole flowering period, or a mix of UVA and UVB for the last 20 days. Day temperature was about 26 °C, with 40 % RH and no added CO₂.

Flower dry weight was 1.6 times higher at the highest light level than at the lowest. Yet "There were also no intensity or UV treatment effects on inflorescence cannabinoid concentrations". The flowers held about 216–249 mg/g THCA in every treatment, and CBD stayed below the detection limit throughout. UV treatment raised THC concentration in the small leaves around the flowers by about 30 % in one treatment, but did not change total THC there either. The study had only three composite samples per treatment, so small effects could be missed. Still, neither stronger light nor UV made this THC cultivar produce CBD.

These findings sit alongside the broader literature summarised by Ahsan et al. (2026), which reports that UV-B can raise cannabinoid accumulation, and that in cell cultures of Cannabis indica, UV raised the expression of the CBDA synthase gene about four-fold. Cell cultures in a flask are not flowering plants, and the review itself calls the link "suggestive". Even where UV does increase cannabinoids, there is no evidence that it switches a plant's chemotype.

Stress

Many growers believe that stress, such as drought, cold nights or physical damage, can push a plant towards one cannabinoid or the other. None of the studies read for this article tested deliberate stress on the THC:CBD ratio, so the belief can be neither confirmed nor dismissed here. The genetic picture suggests what to expect: a plant that lacks a working CBDA synthase cannot make meaningful CBD, however it is treated, because the enzyme is missing, not switched off.

Harvest timing and the ratio

What changes over the flowering weeks

Harvest timing does change the chemistry, but mainly the amount. Ahsan et al. (2026) cite work showing that cannabinoid concentration rises more than five-fold between early and late flower stages. The most detailed week-by-week dataset here comes from Alden's (2025) doctoral dissertation at Clemson University (chapter 2). The study used the CBD-dominant cultivar 'Southern OG', a type III hemp cultivar, in a greenhouse with blackout curtains for a 12-hour day and supplementary LED light. Plants grew in 2.5-litre pots of peat-based medium at 8 plants per m² and about 23 °C. From week 4 of flowering, five plants were harvested each week, up to week 8 in one experiment and week 10 in the other, and the flowers were analysed by a commercial laboratory.

Concentrations changed considerably. In the longer experiment, THC reached a maximum of 0.68 % at week 6 and slowly fell to 0.57 % by week 10; CBD peaked in weeks 6–7 and then declined. Flower mass kept increasing until the last harvest. So the best week for a high concentration was not the best week for the most flower, a trade-off the dissertation explores in detail.

What did not change

The CBD:THC ratio hardly moved. It ranged from 24:1 to 28:1 across all harvests in one experiment and from 21:1 to 24:1 in the other. Harvest week did have a statistically significant effect on the ratio in the second experiment, but there was "not a clear trend". Alden concludes that the ratio is "not impacted by the timing of sampling", and suggests that regulators could use it to tell hemp from drug-type cannabis. That is relevant because THC concentration in this licensed hemp cultivar passed the US legal limit of 0.3 % by week 5 in both experiments, even though the ratio clearly identified it as hemp.

Rodriguez-Morrison et al. (2021) and Alden (2025) both used high-CBD cultivars. Whether the small CBD fraction in a THC-dominant plant shifts over the flowering weeks was not measured in these studies. Nothing here supports the idea that a late harvest turns a THC cultivar into a CBD one.

Trichome colour

Many growers time harvest by trichome colour, from clear through milky to amber; Alden notes that growers often harvest at 10–30 % amber heads. The dissertation describes the evidence for this, and for fixed-calendar harvesting, as "largely anecdotal" and calls for research linking trichome colour to cannabinoid concentration. Trichome watching may help judge maturity, but this research has not shown that it predicts the THC:CBD ratio.

Seed-bank labels and lab tests

A seed packet's "CBD-rich" or "1:1" is a claim about the parent plants, not a measurement of the seed in your hand. Henry et al. (2020) show why both parts of the claim deserve caution: names are sometimes attached to the wrong genetics, and seed lines can be far from uniform. In a 1:1 line, where plants carry one THC-type and one CBD-type version, the offspring can split. Some plants may inherit two THC-type copies, some two CBD-type copies, and some one of each. By the same genetic reasoning, a single pack of a "balanced" cultivar may give plants of quite different chemotypes.

Only a laboratory analysis of the flowers shows the real ratio. The studies here used high-performance liquid chromatography (HPLC), which measures the acid and neutral forms separately. Totals are then calculated as THC = THCA × 0.877 + THC, and likewise for CBD, because the acids lose mass when converted (Rodriguez-Morrison et al., 2021; Alden, 2025). When comparing lab reports, it is worth checking whether they quote "total" values or only the neutral forms. Henry's work also shows that a DNA marker test on a small leaf sample can predict the chemotype long before flowering, but such tests are mainly used by breeders and licensed producers.

Why research and practice seem to disagree

The belief that growing tricks change a plant's character has understandable roots. Growers do see differences between harvests: more or less resin, a different smell, a stronger or weaker effect. The research agrees that amounts and terpenes do change. In Rodriguez-Morrison et al. (2021), total terpenes rose about 1.4-fold with light. What the research does not support is a change of type.

There are also limits to the evidence. The two clearest light studies each used one cultivar in one commercial facility, and only one of them used a THC-dominant plant. The harvest-timing data come from a single hemp cultivar in a greenhouse. The genetic survey covers North American material. The epigenetic ideas in Ahsan et al. (2026) are hypotheses that have not yet been tested in flowering cannabis. And none of these studies measured what the people consuming the flowers experience.

What this means for a 1×1 tent

These are educational takeaways, not instructions.

• Choose the chemotype when you choose the seed. If you want a CBD-dominant or balanced plant, that decision is made at purchase, not in the tent. No light level, UV lamp or harvest week in these studies turned one type into another.
• Grow for quantity and quality with the environment. Good light and a healthy, full flowering period raise the total harvest and its total cannabinoid content. In the light studies, yield rose 1.6-fold to 4.5-fold while concentration stayed flat, so light mainly buys more flowers.
• Expect variation within a seed pack. Especially with "balanced" or hybrid seed, plants from one pack may differ in chemotype. Keeping notes on each plant helps you learn what a line really does.
• Treat labels as claims. Identical genetics are sometimes sold under different names, and plants from one named line can differ genetically.
• Harvest timing is about amount and maturity. In the hemp data, concentrations peaked around weeks 6–7 of a 12-hour day and slowly declined, while the ratio stayed steady. Harvest timing changes how much you get, not which type you have.
• Only a lab test answers the ratio question. Neither appearance, smell nor trichome colour reveals the THC:CBD ratio.

What we know, and what we don't

We know that THCA and CBDA come from the same precursor, CBGA, and are made by two competing enzymes (Ahsan et al., 2026). We know that whether a plant carries working or broken versions of these linked synthase genes largely sets its chemotype, and that DNA markers in this region on chromosome 9 separate CBD-producing from THC-producing plants (Henry et al., 2020). We know that much more light can multiply flower and cannabinoid yield per m² several-fold without changing cannabinoid concentration or the ratio, in a high-CBD cultivar (Rodriguez-Morrison et al., 2021), and that neither stronger light nor UV made a high-THC cultivar produce CBD (Llewellyn et al., 2022). We know that in one hemp cultivar, CBD:THC stayed between about 21:1 and 28:1 at every harvest week while concentrations rose and fell (Alden, 2025). We also know that names and seed lines are less reliable than labels suggest (Henry et al., 2020).

We don't know whether deliberate stresses such as drought, cold or wounding shift the ratio in THC-dominant cultivars; none of the studies here tested them. We don't know whether the small CBD fraction in a THC plant changes with harvest timing, because the week-by-week data come from a CBD cultivar. We don't know how far epigenetic changes, for example in clones kept for many generations, affect cannabinoid production in flowering plants; Ahsan et al. (2026) present these as hypotheses. And the genetic survey covered North American material; how mixed and mislabelled UK and European seed lines are has not been measured here.

The overall picture is simple. The tent decides how much, the genes decide what kind, and only a laboratory can confirm what you actually grew.

Sources

Ahsan, S.M., Keya, S.S., Injamum-Ul-Hoque, M., Howlader, N.C., Hasan, M.R., Haque, M.A., Shekhar, H.U., Choi, H.W. and Rahman, M.M. (2026) 'Epigenome editing for cannabinoid yield: targets, tools, and priorities in Cannabis sativa', International Journal of Molecular Sciences, 27(14), 6299. doi:10.3390/ijms27146299.

Alden, M.J. (2025) Horticultural Strategies to Improve Yield and Efficiency for Controlled-Environment Cannabis Production. PhD dissertation. Clemson University. Available at: https://open.clemson.edu/all_dissertations/4161.

Henry, P., Khatodia, S., Kapoor, K., Gonzales, B., Middleton, A., Hong, K., Hilyard, A., Johnson, S., Allen, D., Chester, Z., Jin, D., Rodriguez Jule, J.C., Wilson, I., Gangola, M., Broome, J., Caplan, D., Adhikary, D., Deyholos, M.K., Morgan, M., Hall, O.W., Guppy, B.J. and Orser, C. (2020) 'A single nucleotide polymorphism assay sheds light on the extent and distribution of genetic diversity, population structure and functional basis of key traits in cultivated north American cannabis', Journal of Cannabis Research, 2, 26. doi:10.1186/s42238-020-00036-y.

Llewellyn, D., Golem, S., Foley, E., Dinka, S., Jones, A.M.P. and Zheng, Y. (2022) 'Indoor grown cannabis yield increased proportionally with light intensity, but ultraviolet radiation did not affect yield or cannabinoid content', Frontiers in Plant Science, 13, 974018. doi:10.3389/fpls.2022.974018.

Rodriguez-Morrison, V., Llewellyn, D. and Zheng, Y. (2021) 'Cannabis yield, potency, and leaf photosynthesis respond differently to increasing light levels in an indoor environment', Frontiers in Plant Science, 12, 646020. doi:10.3389/fpls.2021.646020.