✓ Last fact-checked: November 16, 2026
Update: On November 11, 2026, Cornell and USDA researchers published CannSelect, a new genotyping platform for cannabis breeding, in PLOS ONE. Read the study (PLOS ONE)

Cannabis Genetic Testing Just Got Sharper: What CannSelect Means for Home Growers

Cannabis genetic testing: seedlings in a home grow tent beside a rack of leaf-sample tubes, with a DNA helix rising from one plant
A leaf punch today, a better seed tomorrow.

Cannabis genetic testing just took a real step forward, and this one came out of a university lab, not a seed bank’s marketing team. Last week a Cornell-led team, working with USDA scientists, published a tool called CannSelect that reads targeted spots across the cannabis genome, and in its first test it found the autoflower gene region right where earlier research said it would be. Here’s the grounded answer for home growers: you can’t buy this as a kit, but it’s the kind of tool that makes the seeds you do buy more honest.

Why should someone with one tent in a spare room care? Because every frustration we talk about in the group, the “auto” that never flowered on its own, the pack where half the phenos look like different strains, the surprise herm, starts with genetics nobody checked. Better breeding tools mean fewer of those surprises. Let’s break down what the study did, what it found, and what it didn’t prove.

What is CannSelect, and what did the study actually do?

According to the paper in PLOS ONE, “CannSelect: A high-quality genotyping platform for Cannabis sativa,” published November 11, 2026, the platform uses custom DNA capture probes built on Agilent’s SureSelect system. In plain words: instead of reading a plant’s whole genome, which is slow and expensive, it fishes out a chosen set of spots that tell breeders the most. The lead author is Dustin Wilkerson at Cornell’s School of Integrative Plant Science in Geneva, New York, and the corresponding author is Lawrence Smart, with co-authors from USDA-ARS, Colorado State University, Lancaster University, Organigram and Agilent.

Two kinds of targets went into the design. The first is a genome-wide set of markers (single-letter DNA differences between plants) picked from about 23 million candidates across four public datasets. The second is a list of 148 genes the team cared about, covering disease resistance, stress tolerance, flowering time, plant size, trichome density, cannflavin production, the CBDA synthase gene and sequences found only on the male Y chromosome. The finished version 1.0 has 18,040 unique probes.

The project was funded by the Foundation for Food and Agriculture Research and the New York State Department of Agriculture and Markets, the paper says. It also discloses that one author worked for Agilent during part of the project and another owns a genomics company. That’s normal in plant science, and I like that it’s stated up front. The Marijuana Herald reported the study on November 12.

Can cannabis genetic testing spot an autoflower before it flowers?

That was the first real test, and it passed. The team ran the early version of the platform on 321 plants from ‘Carolina Dream’ and ‘BaOx’ hemp breeding material, grown in a greenhouse and scored simply as day-neutral (autoflowering) or not. Then they ran a genome-wide association scan, which looks for DNA spots that line up with a trait.

The scan lit up chromosome 1 at about 22.47 million base pairs. According to the paper, that’s right inside the window earlier work had mapped for a locus called Autoflower1 (roughly 17.7 to 22.9 Mb on chromosome 1). The paper also notes a second known region, Autoflower2, on chromosome 8. So the tool found the autoflower signal exactly where it should have, which is what you want from a new instrument before you trust it with anything else.

For breeders, that matters a lot. If you can check a seedling’s DNA for the autoflower region in week one, you don’t have to grow out hundreds of plants to find out which ones flower on their own. That’s time, space and power saved, and it’s how “autos closing the gap” keeps happening. I wrote about that gap in Autoflowers Are Closing the Gap, and tools like this are a big reason why.

What this doesn’t prove (yet)

Here’s where I want to be fair. The first version of the platform captured an average of 35.8% of its targets on that test population, which the authors found too loose. They tightened the design, and version 1.0 reached an average of 88.5% when they tested it on 15 hemp accessions from the USDA-ARS germplasm collection. That’s a big jump, but it measures how well the probes grab DNA. The trait mapping in this paper was done with the early version, so the improved one still has to prove itself on real breeding questions.

It also doesn’t give you a “THC gene test” or a herm-proof seed. The design includes Y-chromosome sequences and the CBDA synthase gene, but this paper’s trait result is the autoflower one. Most of the test plants were hemp, not the cultivars in your tent. And it’s a lab platform for breeders and researchers, not something you mail a leaf to tomorrow.

What I do love: the team says the genotyping data and the probe design code are public. That’s the opposite of locked-up “secret genetics.” Open tools help small breeders compete with the big guys, and the small breeders are the ones home growers buy from. It’s also a reminder of why the federal hemp fight matters to us. A lot of this science runs on hemp material, and I covered what the new hemp rules mean for seeds in Hemp Ban and Cannabis Seeds.

From the Grow Room

My take is simple: this is the good kind of cannabis science. It’s curious, it’s checkable, and it points at things growers actually fight with. I’ve said it about light schedules in my 13/11 post and I’ll say it here: curiosity is how cultivation evolves.

Here’s what I’d do with this news as a home grower:

  • Ask your breeder how they select. Do they test for the autoflower trait, sex or chemotype, or do they just grow out and hope? Good breeders will tell you.
  • Log days to flower on every auto. If an “auto” stalls in veg, write it down and tell the breeder. Your notes are field data.
  • Run a small phenohunt before you commit. Pop a few seeds, keep notes, take cuts of the keeper. My guide on cloning with strategic cuts covers that part.
  • Still watch for stress herms. Better genetics help, but environment still matters. Here’s what the research says causes herms.

If you want to try breeding yourself, start with cannabis breeding basics. You don’t need a DNA lab to make good seeds. You need patience, notes and honest selection.

Key Takeaways

  • CannSelect is a new cannabis genotyping platform from Cornell and USDA researchers, published in PLOS ONE on November 11, 2026.
  • In its first test on 321 plants it found the Autoflower1 region on chromosome 1, right where earlier research placed it.
  • The improved version captured 88.5% of its targets on average, but it still has to prove itself on new traits.
  • It’s a breeder’s tool, not a home kit. Ask breeders how they select, and log days to flower on your autos.

Sources

Join the Conversation

Have you ever bought “autoflower” seeds that didn’t act like autos? Would you pay a little more for seeds from a breeder who shows their testing? Tell me in the comments or in the MadManPlant Facebook group.

Want more like this? Join the MadManPlant community and grab the free grow guides.

Stay lifted. Stay learning.

Written by Stefan Walter · Originally published 2026-11-16 · Updated 2026-11-16

author-avatar

About Stefan Walter

Stefan Walter (MadManPlant) is an educator and storyteller focused on the relationship between home cultivation, discipline, and mental wellness. Through lived experience, he translates plant systems into human systems — emphasizing patience, consistency, and intentional growth. His work centers on education, community, and the belief that sustainable outcomes are built, not rushed.

Join 27,000+ Growers: Ask a question or share your progress below.