Deep Learning for Strawberry Canopy Delineation and Biomass Prediction from High-Resolution Images.

Modeling plant canopy biophysical parameters at the individual plant level remains a major challenge. This study presents a workflow for automatic strawberry canopy delineation and biomass prediction from high-resolution images using deep neural networks. High-resolution (5 mm) RGB orthoimages, near-infrared (NIR) orthoimages, and Digital Surface Models (DSM), which were generated by Structure from Motion (SfM), were utilized in this study. Mask R-CNN was applied to the orthoimages of two band combinations (RGB and RGB-NIR) to identify and delineate strawberry plant canopies. The average detection precision rate and recall rate were 97.28% and 99.71% for RGB images and 99.13% and 99.54% for RGB-NIR images, and the mean intersection over union (mIoU) rates for instance segmentation were 98.32% and 98.45% for RGB and RGB-NIR images, respectively. Based on the center of the canopy mask, we imported the cropped RGB, NIR, DSM, and mask images of individual plants to vanilla deep regression models to model canopy leaf area and dry biomass. Two networks (VGG-16 and ResNet-50) were used as the backbone architecture for feature map extraction. The R 2 values of dry biomass models were about 0.76 and 0.79 for the VGG-16 and ResNet-50 networks, respectively. Similarly, the R 2 values of leaf area were 0.82 and 0.84, respectively. The RMSE values were approximately 8.31 and 8.73 g for dry biomass analyzed using the VGG-16 and ResNet-50 networks, respectively. Leaf area RMSE was 0.05 m2 for both networks. This work demonstrates the feasibility of deep learning networks in individual strawberry plant extraction and biomass estimation.

A novel allele of the P-starvation tolerance gene OsPSTOL1 from African rice (Oryza glaberrima Steud) and its distribution in the genus Oryza.

KEY MESSAGE: We have developed allele-specific markers for molecular breeding to transfer the PSTOL1 gene from Kasalath to African mega-varieties, including NERICAs, to improve their tolerance to P-deficient soil. The deficiency of phosphorus (P) in soil is a major problem in Sub-Saharan Africa due to general nutrient depletion and the presence of P-fixing soils. Developing rice cultivars with enhanced P efficiency would, therefore, represent a sustainable strategy to improve the livelihood of resource-poor farmers. Recently the Pup1 locus, a major QTL for tolerance to P deficiency in soil, was successfully narrowed-down to a major gene, the protein kinase OsPSTOL1 (P-starvation tolerance), which was found to be generally absent from modern irrigated rice varieties. Our target is to improve the tolerance of African mega-varieties to P deficiency through marker-assisted introgression of PSTOL1. As a first step, we have determined the Pup1 haplotype and surveyed the presence or absence of PSTOL1 and other genes of the Pup1 locus in African mega-varieties, NERICAs (New Rice for Africa) and their Oryza glaberrima parents. Here, we report the presence of a novel PSTOL1 allele in upland NERICAs that was inherited from the O. glaberrima parent CG14. This allele showed a 35 base-pair substitution when aligned to the Kasalath allele, but maintained a fully conserved kinase domain, and is present in most O. glaberrima accessions evaluated. In-silico and marker analysis indicated that many other genes of the Kasalath Pup1 locus were missing in the O. glaberrima genome, including the dirigent-like gene OsPupK20-2, which was shown to be downstream of PSTOL1. We have developed several allele-specific markers for the use for molecular breeding to transfer the PSTOL1 gene from Kasalath to African mega-varieties, including NERICAs.

The protein kinase Pstol1 from traditional rice confers tolerance of phosphorus deficiency.

As an essential macroelement for all living cells, phosphorus is indispensable in agricultural production systems. Natural phosphorus reserves are limited, and it is therefore important to develop phosphorus-efficient crops. A major quantitative trait locus for phosphorus-deficiency tolerance, Pup1, was identified in the traditional aus-type rice variety Kasalath about a decade ago. However, its functional mechanism remained elusive until the locus was sequenced, showing the presence of a Pup1-specific protein kinase gene, which we have named phosphorus-starvation tolerance 1 (PSTOL1). This gene is absent from the rice reference genome and other phosphorus-starvation-intolerant modern varieties. Here we show that overexpression of PSTOL1 in such varieties significantly enhances grain yield in phosphorus-deficient soil. Further analyses show that PSTOL1 acts as an enhancer of early root growth, thereby enabling plants to acquire more phosphorus and other nutrients. The absence of PSTOL1 and other genes-for example, the submergence-tolerance gene SUB1A-from modern rice varieties underlines the importance of conserving and exploring traditional germplasm. Introgression of this quantitative trait locus into locally adapted rice varieties in Asia and Africa is expected to considerably enhance productivity under low phosphorus conditions.

Developing rice with high yield under phosphorus deficiency: Pup1 sequence to application.

The major quantitative trait locus (QTL) Phosphorus uptake1 (Pup1) confers tolerance of phosphorus deficiency in soil and is currently one of the most promising QTLs for the development of tolerant rice (Oryza sativa) varieties. To facilitate targeted introgression of Pup1 into intolerant varieties, the gene models predicted in the Pup1 region in the donor variety Kasalath were used to develop gene-based molecular markers that are evenly distributed over the fine-mapped 278-kb QTL region. To validate the gene models and optimize the markers, gene expression analyses and partial allelic sequencing were conducted. The markers were tested in more than 80 diverse rice accessions revealing three main groups with different Pup1 allele constitution. Accessions with tolerant (group I) and intolerant (group III) Pup1 alleles were distinguished from genotypes with Kasalath alleles at some of the analyzed loci (partial Pup1; group II). A germplasm survey additionally confirmed earlier data showing that Pup1 is largely absent from irrigated rice varieties but conserved in varieties and breeding lines adapted to drought-prone environments. A core set of Pup1 markers has been defined, and sequence polymorphisms suitable for single-nucleotide polymorphism marker development for high-throughput genotyping were identified. Following a marker-assisted backcrossing approach, Pup1 was introgressed into two irrigated rice varieties and three Indonesian upland varieties. First phenotypic evaluations of the introgression lines suggest that Pup1 is effective in different genetic backgrounds and environments and that it has the potential to significantly enhance grain yield under field conditions.