Skip to main content
ARS Home » Southeast Area » Stuttgart, Arkansas » Dale Bumpers National Rice Research Center » Research » Publications at this Location » Publication #371777

Research Project: Gene Discovery and Crop Design for Current and New Rice Management Practices and Market Opportunities

Location: Dale Bumpers National Rice Research Center

Title: Genomic prediction and bayesian network analysis of multiple root architecture traits in rice

item Sharma, Santosh
item Gealy, David
item Pinson, Shannon
item Edwards, Jeremy

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 1/24/2020
Publication Date: 2/10/2020
Citation: Sharma, S., Gealy, D.R., Pinson, S.R., Edwards, J. 2020. Genomic prediction and bayesian network analysis of multiple root architecture traits in rice. Poster presentation. Arkansas Bioinformatics Consortium, Little Rock, Arkansas. February 10-11, 2020.

Interpretive Summary:

Technical Abstract: Root system architecture (RSA) is a crucial factor in resource acquisition and plant productivity. High-throughput root image analysis systems can expedite the genetic study of RSA. Roots are difficult to phenotype in the field, thus new tools for predicting phenotype from genotype are particularly valuable for plant breeders aiming to improve RSA. The objectives of this study were (1) to discover quantitative trait loci (QTL) for RSA traits in rice, (2) to evaluate the accuracy of genomic prediction methods for RSA traits, and (3) to model the network of interactions among the RSA traits, RSA QTL, and their interactions with above-ground agronomic traits. A recombinant inbred line (RIL) population derived from parents with contrasting RSA (PI312777 x Katy) was genotyped with the 7K SNPs (C7AIR) and phenotyped for RSA on seedlings grown on agar plates and agronomic traits on plants in the field. QTL mapping and genomic prediction were used to investigate the RSA traits and their interactions with agronomic traits. The genomic prediction methods evaluated included rrBLUP, GBLUP and BRR. Multi-trait QTL and trait interactions were modeled using the “Bnlearn” Bayesian network (BN) analysis package in R. Genomic prediction explained a higher proportion of genetic variance (56% for grain yield, 32% root surface area, 58% root hair surface area/total root area) for most quantitative traits compared to multi-QTL models (26% grain yield, 30% root surface area, 34% root hair surface area/total root area). Multi-trait BN analysis was found to improve genomic prediction (3-20%). The BN analysis confirmed QTL identified by single-trait QTL analysis and identified new QTL including one proximal to the pleiotropic gene NAL1 on chromosome four affecting root length, leaf width and chlorophyll content. This study demonstrates the effectiveness of genomic prediction and multi-trait Bayesian network methods to model RSA traits and their above-ground trait relationships.