Adu, Michael Osei
(2014)
Variations in root system architecture and root growth dynamics of Brassica rapa genotypes using a new scanner-based phenotyping system.
PhD thesis, University of Nottingham.
Abstract
There is a need to breed for root systems architectures (RSAs) that optimise soil resource acquisition. This requires high resolution and high-throughput quantification of RSA in as natural an environment as possible. Current imaging techniques are limited by cost, reproducibility, throughput and complexity. This thesis describes (1) the construction of a low cost, high-resolution, root phenotyping platform that requires no sophisticated equipment which is adaptable to most laboratory and glasshouse environments and (2) its application to quantify environmental and temporal variation in RSA between genotypes of Brassica rapa L.
The high resolution root phenotyping system (HRP) that was constructed employed 24 scanners and could screen up to 72 individual plants at any time, with the possibility of capturing thousands of root images daily depending on the operational number of scanners and scanning periodicity. Plants were supplied with a complete nutrient solution through the wick of a germination paper. Images of RSA were acquired automatically, over extended periods, using multiple scanners controlled by customised software. The RSA data was used to validate a mechanistic model and mixed effects models were used to describe the sources of variation in traits contributing to RSA. Plants were also grown in rhizoboxes and under varying concentrations of P ([P]ext).
Broad-sense heritability (H2), was highest (≥ 0.70) for shoot biomass, length of primary roots (PRs), number of lateral roots (LRs). Coefficients of variation in RSA traits within a genotype were large and ranged between 5 and 103%. It was found that between 4 and 48 replicates were needed to detect a significant difference (95% CI, 50% difference between trait means). Significant differences were found between genotypes in root traits with strong positive correlations among RSA traits and between biomass and RSA traits. Principal component analyses identified 5 significant axes of variation, accounting for approximately 86 and 78% of the variation in the genotypes on paper and soil substrates, respectively. Cluster analysis of the genotypes produced 5 discrete groups. Genotypes with more or less shoot biomass or with bigger or smaller RSA could be distinguished.
A density-based 2D model reproduced experimental results accurately by simulating PR length and total length of LRs. Mixed-effects statistical models demonstrated that root traits show temporal variations of various types with significant effects of genotype. All genotypes followed a similar growth pattern with time, but differed in their maximum total root length (TRL), primary root length (PRL) and LR growth. A 3-parameter logistic model satisfactorily described TRL and PRL when genotypes were grown on both paper and soil substrates. On paper substrate, TRL required only a single, random-effect parameter (asymptote), describing maximum TRL. On soil substrate, TRL required two random-effects parameters, asymptote and inflection, describing maximum TRL and time at which ½ of maximum TRL occurs, respectively. Primary root length on both paper and soil substrates required only a single, random-effect parameter, describing maximum PRL. The growth rate of LRs of all ages followed a quadratic function and required only a single, random-effect parameter, describing maximum growth rate.
There was variation in specific RSA traits and plasticity in response to [P]ext among genotypes. Length of the apical un-branched zone of the PR increased with increasing [P]ext. Total root length, total LR length and number of LRs was positively correlated with total plant tissue P concentration at low [P]ext but not at high [P]ext. Paper substrate was more suitable for screening seedling root traits but root phenotypes must be validated in situ in the field or in soil media because some differences were evident between data observed on paper and soil substrates.
Scanner-based phenotyping of RSA provides economical means of studying the mechanisms underlying the plant-soil interactions and can be used to quantify environmental and temporal variation in traits contributing to RSA. The HRP system can be extended to screen the large populations required for breeding for efficient resource acquisition. The necessity for high replication and time-consuming image analysis could however limit throughput in the phenotyping system.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
