RESUMO
Introduction: Genotyping large-scale gene bank collections requires an appropriate sampling strategy to represent the diversity within and between accessions. Methods: A panel of 44 common bean (Phaseolus vulgaris L.) landraces from the Alliance Bioversity and The Alliance of Bioversity International and the International Center for Tropical Agriculture (CIAT) gene bank was genotyped with DArTseq using three sampling strategies: a single plant per accession, 25 individual plants per accession jointly analyzed after genotyping (in silico-pool), and by pooling tissue from 25 individual plants per accession (seq-pool). Sampling strategies were compared to assess the technical aspects of the samples, the marker information content, and the genetic composition of the panel. Results: The seq-pool strategy resulted in more consistent DNA libraries for quality and call rate, although with fewer polymorphic markers (6,142 single-nucleotide polymorphisms) than the in silico-pool (14,074) or the single plant sets (6,555). Estimates of allele frequencies by seq-pool and in silico-pool genotyping were consistent, but the results suggest that the difference between pools depends on population heterogeneity. Principal coordinate analysis, hierarchical clustering, and the estimation of admixture coefficients derived from a single plant, in silico-pool, and seq-pool successfully identified the well-known structure of Andean and Mesoamerican gene pools of P. vulgaris across all datasets. Conclusion: In conclusion, seq-pool proved to be a viable approach for characterizing common bean germplasm compared to genotyping individual plants separately by balancing genotyping effort and costs. This study provides insights and serves as a valuable guide for gene bank researchers embarking on genotyping initiatives to characterize their collections. It aids curators in effectively managing the collections and facilitates marker-trait association studies, enabling the identification of candidate markers for key traits.
RESUMO
Crop diversity conserved in genebanks facilitates the development of superior varieties, improving yields, nutrition, adaptation to climate change and resilience against pests and diseases. Cassava (Manihot esculenta) plays a vital role in providing carbohydrates to approximately 500 million people in Africa and other continents. The International Center for Tropical Agriculture (CIAT) conserves the largest global cassava collection, housing 5,963 accessions of cultivated cassava and wild relatives within its genebank. Efficient genebank management requires identifying and eliminating genetic redundancy within collections. In this study, we optimized the identification of genetic redundancy in CIAT's cassava genebank, applying empirical distance thresholds, and using two types of molecular markers (single-nucleotide polymorphism (SNP) and SilicoDArT) on 5,302 Manihot esculenta accessions. A series of quality filters were applied to select the most informative and high-quality markers and to exclude low-quality DNA samples. The analysis identified a total of 2,518 and 2,526 (47 percent) distinct genotypes represented by 1 to 87 accessions each, using SNP or SilicoDArT markers, respectively. A total of 2,776 (SNP) and 2,785 (SilicoDArT) accessions were part of accession clusters with up to 87 accessions. Comparing passport and historical characterization data, such as pulp color and leaf characteristic, we reviewed clusters of genetically redundant accessions. This study provides valuable guidance to genebank curators in defining minimum genetic-distance thresholds to assess redundancy within collections. It aids in identifying a subset of genetically distinct accessions, prioritizing collection management activities such as cryopreservation and provides insights for follow-up studies in the field, potentially leading to removal of duplicate accessions.
RESUMO
BACKGROUND: Maize cobs are an important component of crop yield that exhibit a high diversity in size, shape and color in native landraces and modern varieties. Various phenotyping approaches were developed to measure maize cob parameters in a high throughput fashion. More recently, deep learning methods like convolutional neural networks (CNNs) became available and were shown to be highly useful for high-throughput plant phenotyping. We aimed at comparing classical image segmentation with deep learning methods for maize cob image segmentation and phenotyping using a large image dataset of native maize landrace diversity from Peru. RESULTS: Comparison of three image analysis methods showed that a Mask R-CNN trained on a diverse set of maize cob images was highly superior to classical image analysis using the Felzenszwalb-Huttenlocher algorithm and a Window-based CNN due to its robustness to image quality and object segmentation accuracy ([Formula: see text]). We integrated Mask R-CNN into a high-throughput pipeline to segment both maize cobs and rulers in images and perform an automated quantitative analysis of eight phenotypic traits, including diameter, length, ellipticity, asymmetry, aspect ratio and average values of red, green and blue color channels for cob color. Statistical analysis identified key training parameters for efficient iterative model updating. We also show that a small number of 10-20 images is sufficient to update the initial Mask R-CNN model to process new types of cob images. To demonstrate an application of the pipeline we analyzed phenotypic variation in 19,867 maize cobs extracted from 3449 images of 2484 accessions from the maize genebank of Peru to identify phenotypically homogeneous and heterogeneous genebank accessions using multivariate clustering. CONCLUSIONS: Single Mask R-CNN model and associated analysis pipeline are widely applicable tools for maize cob phenotyping in contexts like genebank phenomics or plant breeding.
