Skip to main content
ARS Home » Northeast Area » Ithaca, New York » Robert W. Holley Center for Agriculture & Health » Plant, Soil and Nutrition Research » Research » Publications at this Location » Publication #358263

Research Project: Improving Crop Efficiency Using Genomic Diversity and Computational Modeling

Location: Plant, Soil and Nutrition Research

Title: Transcriptome-wide association supplements genome-wide association in Zea mays

item KREMLING, KARL - Cornell University
item DIEPENBROCK, CHRISTINE - Cornell University
item GORE, MICHAEL - Cornell University
item Buckler, Edward - Ed
item BANDILLO, NONOY - Cornell University

Submitted to: Genes, Genomes, Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/18/2019
Publication Date: 9/1/2019
Citation: Kremling, K., Diepenbrock, C., Gore, M., Buckler IV, E.S., Bandillo, N. 2019. Transcriptome-wide association supplements genome-wide association in Zea mays. Genes, Genomes, Genetics. 9(9):3023-3033.

Interpretive Summary: To improve the cultivation of plants, it’s essential to understand how genes (or genotype) affect the physical characteristics of the plant (its phenotype). In the past, most research focused on identifying genetic markers (a short sequence of DNA used to identify a chromosome) and how those markers are linked to gene of interest to determine and better understand the phenotype. However, the connection between genotype and phenotype is more complex and finding a gene responsible for phenotype is like finding a needle in a haystack. Further, a map of genetic markers does not account for all the phenotypic outcomes and some portion of phenotypic variation are left explained. Using gene expression endophenotyping (an intermediate level of biological organization that exist between DNA and observed phenotypic outcomes) applied in a diverse panel of maize, we revealed that about half of the functional variation for increasing yield (agronomic traits) and quality traits (carotenoid, tocochromanol) in maize are regulatory. We found that the variability in gene expression, in combination with nucleotide variation, can help improve the capacity to narrow down and identify the causal variants underlying changes in phenotype. The result improves not only the capacity to link genes to phenotypes, but also illustrates the importance of regulation for phenotype. Plant breeders are always looking for useful subtle variation to breed new high-yielding varieties with increased resistance to pest and diseases and nutritional quality, best suited for landscapes and climate at farmer’s field. Breeding crops can be more efficient if researchers could rank and find what types of molecular changes in the crop genome underlie phenotypic variation. Our results can help on revealing DNA-level and gene-level knowledge that can be used for breeding cultivated crops around the globe.

Technical Abstract: Discovery of variation that underlies quantitative traits remains central to the genetic improvement of agricultural species. To date, the search has mostly relied on associations between genetic markers and observed phenotypes. However, multiple levels of biological organization exist between the ultimate driver of phenotype and terminal observed phenotypic outcomes, enabling trait dissection to be done between intermediate levels of biological organization (hereafter, endophenotypes). Here, we illustrate the power of using gene expression endophenotypes measured in a large gene expression resource (299 individuals, seven tissues) collected from the Goodman diversity panel in Zea mays. Using single-tissue- and multi-tissue-based transcriptome-wide association studies (TWAS), we revealed that about half of the functional variation for carotenoid, tocochromanol and agronomic traits are regulatory. Comparing the efficacy of TWAS with genome-wide association studies (GWAS) and an ensemble approach that combines both GWAS and TWAS, we demonstrated that results of TWAS in combination with GWAS increase the power to detect known genes and aid in prioritizing likely causal genes. Using a variance partitioning approach in the independent maize Nested Association Mapping (NAM) population, we also showed that the most strongly associated genes identified by combining GWAS and TWAS explain more heritable variance, beating the heritability captured by the set of random genes for a majority of the traits. This improves not only the ability to link genes to phenotypes, but also highlights the phenotypic consequences of regulatory variation in plants.