Reducing herbivore damage using biodiversity instead of insecticide

One of the key challenges of modern society lies in reconciling food security and the preservation of the environment and biodiversity. Pests and diseases pose serious threats to crops, making chemical agents like pesticides crucial in agriculture. However, pesticides can reduce the biodiversity of insects. “In this context, associative resistance could be a new method to ensure food production while preserving biodiversity,” says Kentaro Shimizu, Director of the Department of Evolutionary Biology and Environmental Studies at the University of Zürich (UZH) in Switzerland.

But which combinations of plant genotypes should be planted together to effectively resist pests and diseases? For example, if one selects two genotypes from a total of 199 genotypes, there are 19,701 possible combinations. Researchers at UZH have now developed new methods of genomic prediction using a physics model to analyze interactions between individuals at the genetic level.

Extensive fieldwork in the research garden
 

First, the researchers conducted large-scale plant cultivation experiments over two years in open fields in Switzerland and in Japan. For the 199 genotypes of the model plant Arabidopsis thaliana collected from around the world, genomic DNA information was already available.

The researchers randomly mixed and planted 30+ individuals from each of the genotypes for a total of 6,400 plants. “To count 52,707 insects on 6,400 plants, the senior researcher Yasuhiro Sato spent months in the research garden. This amazing dataset, collected by taking advantage of the research garden at the Irchel campus in Switzerland, was the key to this study,” says Kentaro Shimizu.

Until now, there were no methods to analyse which genomic regions underpin interactions such as associative resistance between neighbouring plant individuals. Yasuhiro Sato and his team therefore developed a new analytical method called Neighbour GWAS. This method applies a model used in physics to analyse interactions between magnets to the interactions between neighbouring plant individuals. It examines how herbivore damage is affected when individuals with specific genetic DNA sequences are adjacent, based on the actual results from field experiments.

Up to 25 per cent less herbivore damage
 

From the analysis using this new method, it was shown that numerous genes are involved in interactions with surrounding individuals. Using a machine learning method, the researchers were able to predict herbivore damage and identified beneficial genotype combinations for which associative resistance was predicted.

The research team conducted another large-scale field experiment over two years, planting around 2,000 plant individuals in pairs of genotypes for which three different levels of associative resistance were predicted. This experiment revealed that, compared to planting a single genotype, mixing two genotypes reduced herbivore damage by 24.8 and 22.7 per cent, respectively, for the highest and second-highest associative resistance level.

“From the perspective of basic research, this can be seen as a landmark in the study of interactions between plant individuals,” says Kentaro Shimizu. “It highlights the importance of biodiversity in two ways. First, the genetic diversity of crops themselves can reduce pest damage. Second, reducing the use of pesticides in agricultural settings can contribute to the conservation of biodiversity, including that of insects.”

Meta-analyses with Bernhard Schmid as co-author have shown that in crops such as wheat or rice, yield increases from 4–16 per cent are achieved if random genotypes are mixed in the field.

According to Shimizu, for these important agricultural plant species whose genomes are known, the new method makes it possible to predict mixtures of specific plant genotypes that maximize associative resistance, thus increasing yield even further while at the same time saving on pesticide use.

(UZH/wi)

Reference:
Yasuhiro Sato, Rie Shimizu-Inatsugi, Kazuya Takeda, Bernhard Schmid, Atsushi J. Nagano, Kentaro K. Shimizu. Reducing herbivory in mixed planting by genomic prediction of neighbor effects in the field. 30 September 2024. Nature Communications.