Work from an international private-public research consortium promises to bring improved canola varieties to farmers’ fields in record time, thanks to the marriage of genomics and artificial intelligence.
“Ultimately, it’s a benefit to producers. They’ll see a faster turnaround in variety production, delivery of new varieties, better varieties,” said Andrew Sharpe, a researcher at the Global Institute for Food Security in Saskatoon.
The canola “pan-genome” project looked at a dozen varieties of Brassica napus submitted by industry partners. These include Bayer, Corteva Agriscience, Nutrien Ag Solutions, and Nuseed.
Sharpe led the project, along with Isobel Parkin from Agriculture Canada’s Saskatoon Research Centre. Both are experts in genomics and bioinformatics, that is, the application of advanced computing technology to biological questions.
While much is already known about canola genetics, this project aimed to sequence the genomes of several varieties at once. This allows them to be looked at side-by-side to compare where they are different.
For example, if one canola variety is known to have good disease resistance, it can be compared against a disease-susceptible variety to quickly identify which genes may be responsible. It also allows easy comparison among quite different varieties — both spring and winter canola were among the dozen sequenced.
That the project, launched in 2019, was completed in just two years is testament to the contributions of industry partner NRGene. By combining advanced artificial intelligence computing systems with extensive genomic databases, it has drastically cut the time required to decode the genetics of any crop.
Plant genomes can be complex. Animals, including people, have diploid genomes, that is, half their genes come from each parent. Plants can be polyploid, containing all the genes of all parents. Bread wheat, for example, is hexaploid, containing the genomes of three different ancestors. Canola has two.
“Essentially it’s a cabbage genome coming together with a turnip genome and that gives you napus, Brassica napus, which is canola,” Sharpe said.
He explained the genus Brassica is the “dog of the plant world,” in that changing just a few genes can radically alter what the plant looks like. Cabbage, Brussels sprouts, mustard, cauliflower, broccoli, and rutabagas are all Brassicas.
Sharpe said this sort of rapid multi-genome sequencing was not possible a few years ago. The job required breaking apart the 12 genomes into manageable pieces, sequencing them, then stitching them back together in the proper order.
“What you end up with when you sequence the genome is you take the DNA, which is very long fragments, and you shatter it into smaller fragments,” he said. “So you end up with what is essentially a billion-piece jigsaw puzzle.”
To make things more challenging, many of the genetic sequences are quite similar, Sharpe said. The puzzle is mostly blue sky, with a few clouds. This is where NRGene’s technology came in.
“The challenge is to stitch everything back together again, the genome sequence,” he said. “That is what NRGene was very successfully doing… they brought out algorithms, novel algorithms that could do this really quite routinely.”
Masood Rizvi is head of genomics for NRGene’s Saskatoon office. He said their technology is algorithm based, using artificial intelligence to crunch the vast amounts of genomics data into meaningful information that can be applied in breeding programs.
It’s speeded up the process considerably. Rizvi said it usually takes seven to 10 years to introduce a useful trait and release it in a new variety.
“With our technology, we cut down the cycle by half,” he said. “It means if you are spending 10 years to bring in those genes, you just need only five years.”
The reason for this is if the breeders know the genetic markers for, say, early-season vigour in winter canola, they can test the lines before they get to the field to make sure they have this trait. Once they get into field trials, breeders backcross plants with their parent lines to keep the traits they want and get rid of the rest. Winter canola, for example, cannot survive Canadian winters, so much of its genome would be undesirable in spring varieties.
Rizvi said with each backcross, the lines can be checked to make sure they have the desired genes, something called genomic selection. This means dramatic reductions in field testing and cost. A company or research organization might take 10,000 lines into field trials and wait six to nine months to grow them up and get results.
“You screen those lines with genomic selection and do a prediction. Out of the prediction, you only take 500 lines to the field,” he said. “That’s the advantage for field screening, because genotyping is very cheap as compared to what we grow in the field.”
NRGene, based in Israel, is a relative newcomer to Canada, having arrived at Saskatoon’s Innovation Place in 2020. Rizvi said they plan to establish their own genotyping laboratories and offer their services to the crop development community. A goal is to reduce genotyping costs to make it more available not only for canola, but for other crops, to benefit crop developers and ultimately, farmers.
“Once this all really speeds up, it’s a boon for everybody in the Canadian economy,” Rizvi said.