Articles
Identification of QTLs for Japanese plum fruit color, size and shape using a computer vision phenotyping approach
Article number
1450_47
Pages
357 – 364
Language
English
Abstract
The mapping of quantitative trait loci (QTLs) for fruit quality traits requires the phenotyping of fruits harvested from a large number of genotypes, a laborious and time-consuming task that increases exponentially with the number of traits, individuals, and fruits evaluated.
We have developed a pipeline to high-throughput phenotype color, size and shape of fruits, which relies on a machine-learning algorithm that detects plums from standardized images before their data collection.
The objective of this study was to validate this pipeline for its use in genomic studies aimed at detecting QTLs for these traits.
For this, we used this methodology to image-phenotype 675 plum fruits from a test panel of 153 trees that have been genotyped using double digest restriction-site associated DNA sequencing (ddRAD-seq) technology.
Phenotypic data distribution, correlation and principal component analysis (PCA) results indicated that the measures were able to differentiate fruits according to their color, size and shape.
Finally, we conducted a genome-wide association study (GWAS) in the test panel using a mixed linear model (MLM) and the obtained phenotypic data.
Several QTLs were detected, most of which were newly discovered and others were consistent with previous studies, such as the Hue color in LG3 and size and aspect ratio in LG7. We conclude that this phenotyping pipeline can effectively be used in future GWAS and QTL analyses in plum, and potentially other types of fruits.
We have developed a pipeline to high-throughput phenotype color, size and shape of fruits, which relies on a machine-learning algorithm that detects plums from standardized images before their data collection.
The objective of this study was to validate this pipeline for its use in genomic studies aimed at detecting QTLs for these traits.
For this, we used this methodology to image-phenotype 675 plum fruits from a test panel of 153 trees that have been genotyped using double digest restriction-site associated DNA sequencing (ddRAD-seq) technology.
Phenotypic data distribution, correlation and principal component analysis (PCA) results indicated that the measures were able to differentiate fruits according to their color, size and shape.
Finally, we conducted a genome-wide association study (GWAS) in the test panel using a mixed linear model (MLM) and the obtained phenotypic data.
Several QTLs were detected, most of which were newly discovered and others were consistent with previous studies, such as the Hue color in LG3 and size and aspect ratio in LG7. We conclude that this phenotyping pipeline can effectively be used in future GWAS and QTL analyses in plum, and potentially other types of fruits.
Publication
Authors
A. Fiol, S. Ahumada, M. Osorio, P. Ballesta, R. Infante, I. Pacheco
Keywords
computer vision phenotyping, QTL, GWAS, fruit color, size, shape
Online Articles (59)