Walnut pellicle color is a key quality attribute that drives consumer preference and walnut sales. For the first time a high-throughput, computer vision-based phenotyping platform using a custom algorithm to quantitatively score each walnut pellicle in L* a* b* color space was deployed at large-scale. This was compared to traditional qualitative scoring by eye and was used to dissect the genetics of pellicle pigmentation. Progeny from both a bi-parental population of 168 trees (‘Chandler’ x ‘Idaho’) and a genome-wide association (GWAS) with 528 trees of the UC Davis Walnut Improvement Program were analyzed. Color phenotypes were found to have overlapping regions in the ‘Chandler’ genetic map on Chr01 suggesting complex genetic control. In the GWAS population, multiple, small effect QTL across Chr01, Chr07, Chr08, Chr09, Chr10, Chr12 and Chr13 were discovered. Marker trait associations were co-localized with QTL mapping on Chr01, Chr10, Chr14, and Chr16. Putative candidate genes controlling walnut pellicle pigmentation were postulated.

核桃皮颜色是决定消费者喜好和核桃销售的关键质量属性。首次大规模部署了使用定制算法对L * a * b *颜色空间中的每个核桃薄膜进行定量打分的高吞吐量，基于计算机视觉的表型平台。将其与传统的眼睛定性评分进行比较，并用于剖析防护膜色素沉着的遗传学。从168棵（“钱德勒”两个双父母人口后代X分析了“爱达荷州”和UC戴维斯核桃改良计划的528棵树的全基因组关联（GWAS）。发现颜色表型在Chr01的“钱德勒”遗传图中具有重叠区域，表明复杂的遗传控制。在GWAS人群中，发现了跨Chr01，Chr07，Chr08，Chr09，Chr10，Chr12和Chr13的多个小效应QTL。标记性状关联与QTL映射在Chr01，Chr10，Chr14和Chr16上共定位。假定控制核桃薄膜色素沉着的推定候选基因。