GWAS, MWAS and mGWAS provide insights into precision agriculture based on genotype-dependent microbial effects in foxtail millet

Plant growth and yield are complex traits shaped by both genetic and environmental factors. Genome-wide association studies (GWAS) have been instrumental in identifying genetic loci influencing these traits in various crops. However, plant genotype alone often fails to fully explain observed trait variations. This limitation highlights the importance of considering the plant's interaction with its surrounding environment, specifically the rhizosphere microbiome. The rhizosphere, the soil region immediately surrounding plant roots, is a complex ecosystem teeming with microorganisms that can significantly impact plant health, growth, and yield. Plant genotypes can influence the composition and function of the rhizosphere microbiome, creating an indirect pathway through which host genes modify phenotypic plasticity. Beneficial microbes can alleviate stress, regulate development, and enhance defense responses, potentially offering opportunities for sustainable agricultural improvement. Understanding the intricate interplay between plant genotype, rhizosphere microbiota, and agronomic traits is crucial for developing precision agriculture strategies. Microbiome-wide association studies (MWAS) and microbiome genome-wide association studies (mGWAS) offer powerful approaches to dissect these complex relationships. This study focuses on foxtail millet (*Setaria italica*), an important crop in arid and semi-arid regions, to investigate the combined effects of host genetics and rhizosphere microbiota on key agronomic traits. Foxtail millet provides a suitable model due to its drought tolerance and well-established genetic resources. The study aims to identify key microbial biomarkers associated with growth and yield, validate their effects, and uncover the underlying genetic mechanisms driving genotype-microbiota interactions.

Previous research has successfully used GWAS to identify genes associated with various traits in crops such as maize, rice, sorghum, cotton, and soybean. However, the genetic basis of complex yield traits remains elusive, hindering efforts to improve crop yields through genetic engineering. Studies have shown that the plant genotype can significantly alter the composition of the root microbiome. This interaction influences plant phenotypic plasticity, particularly in response to environmental stress. The root microbiome can mitigate environmental stress, influence plant development, and modulate defense responses. Identifying growth-promoting bacteria and understanding their interactions with the host plant is vital for sustainable agricultural practices. Plant nutrient starvation and immunity significantly affect plant-microbe interactions. MWAS has been applied to human cohorts to identify gut microbial markers for complex traits like obesity and type 2 diabetes. However, its application in plant systems remains relatively limited. While GWAS has revealed some key loci in foxtail millet related to flowering time and disease resistance, loci associated with plant growth and yield are less well understood. A previous study explored the correlation between foxtail millet root zone microbiota and yield traits, hinting at the potential for agricultural improvement through microbiota modification.

This study employed a multi-faceted approach integrating GWAS, MWAS, and mGWAS to comprehensively analyze the relationship between foxtail millet genotype, rhizosphere microbiota, and agronomic traits. A total of 827 foxtail millet cultivars were collected from China and subjected to whole-genome sequencing to identify single nucleotide polymorphisms (SNPs). After rigorous quality control, 161,562 SNPs were retained for analysis. These cultivars were planted in a field trial, and 12 agronomic traits (6 growth traits and 6 yield traits) were measured. GWAS was performed using a linear mixed model to identify SNPs associated with these traits. The rhizosphere microbiome was characterized using 16S rRNA gene sequencing. The data was used in MWAS to identify microbial operational taxonomic units (OTUs) correlated with the agronomic traits. Linear regression models were used to quantify the individual contributions of genetic variations (SNPs) and microbial variations (OTUs) to the phenotypic variations. Mixed linear models were used to investigate the combined effects of SNPs and OTUs. To validate the predicted effects of microbial markers, bacterial strains were isolated from the rhizosphere and their effects on foxtail millet growth were tested in plate and sterilized soil experiments. mGWAS was conducted to identify genetic loci correlated with rhizoplane microbial abundance. RNA sequencing was performed to investigate the expression patterns of candidate genes in response to inoculation with selected bacterial strains. Finally, experiments were conducted on different genotypes to validate the genotype-dependent effects of microbial strains on plant growth.

GWAS identified 86 significant SNP loci associated with 10 of the 12 agronomic traits. Candidate genes near these loci were involved in growth and development regulation, drought stress response, plant immunity, and nutrient uptake. Linear regression models showed that genetic markers alone could explain a substantial portion of the variance in the six key traits (TSLW, MSPD, MSW, MSPW, PGW, and MSPL). MWAS identified 257 marker OTUs significantly associated with the six traits. These OTUs represented diverse bacterial taxa. The combined use of SNP and OTU markers in mixed linear models significantly improved the prediction accuracy for all six traits compared to using only SNPs or OTUs. Experiments with isolated bacterial strains confirmed their growth-promoting or suppressing effects on foxtail millet. mGWAS revealed that the abundance of many marker OTUs was associated with host genetic variations. Plant immune genes, such as *FLS2* and *bHLH35*, were strongly associated with the abundance of specific microbial taxa. Experiments showed that the effects of bacterial strains on plant growth were genotype-dependent; certain strains promoted growth in some genotypes but not others.

This study demonstrates the power of integrating GWAS, MWAS, and mGWAS to unravel the complex interplay between plant genotype, rhizosphere microbiota, and agronomic traits. While GWAS alone provides insights into the genetic architecture of traits, incorporating microbiome data reveals additional layers of complexity. The findings suggest that rhizosphere microbiota play a significant role in shaping foxtail millet growth and yield. The identified microbial biomarkers can inform the development of precision microbiome management strategies to improve crop yields. The genotype-dependent effects of microbial strains highlight the importance of considering host genetics when developing microbiome-based agricultural interventions. The study also sheds light on the genetic mechanisms underlying genotype-microbiota interactions. The observed associations between plant immune genes and specific microbial taxa suggest a potential for manipulating plant immunity to engineer desired microbiome compositions. Future research could focus on exploring the specific mechanisms through which these interactions influence plant phenotypes and translating these findings into practical agricultural applications.

This study provides a comprehensive analysis of the interplay between foxtail millet genotype, rhizosphere microbiota, and agronomic traits. The findings underscore the importance of integrating host genetics and microbiome data to fully understand and improve crop productivity. The identification of key microbial biomarkers and the demonstration of genotype-dependent effects pave the way for developing precision microbiome management strategies for enhancing crop yields. Future research should investigate the underlying molecular mechanisms of these genotype-microbiota interactions, leading to the development of targeted approaches to engineer high-yielding and stress-resilient cultivars. This approach holds great promise for creating more sustainable and efficient agricultural systems.

The study was conducted in a single environment, limiting the generalizability of the findings to other environments. The use of 16S rRNA gene sequencing to characterize the microbiome might have missed some aspects of microbial diversity and function. The validation experiments involved a limited number of bacterial strains, which might not fully represent the diversity of the foxtail millet rhizosphere microbiome. Future studies should address these limitations by expanding the geographical scope, employing more comprehensive microbiome profiling techniques, and expanding the range of bacterial strains included in validation experiments.