Amplicon sequencing (AmpSeq) is a practical, intuitive strategy with a semi-automated computational pipeline for analysis of highly multiplexed PCR-derived sequences. This genotyping platform is particularly cost-effective when multiplexing 96 or more samples with a few amplicons up to thousands of amplicons. Amplicons can target from a single nucleotide to the upper limit of the sequencing platform. The flexibility of AmpSeq’s wet lab methods make it a tool of broad interest for diverse species, and AmpSeq excels in flexibility, high-throughput, low-cost, accuracy, and semi-automated analysis. Here we provide an open science framework procedure to output data out of an AmpSeq project, with an emphasis on the bioinformatics pipeline to generate SNPs, haplotypes and presence/absence variants in a set of diverse genotypes. Open-access tutorial datasets with actual data and a containerization open source software instance are provided to enable training in each of these genotyping applications. The pipelines presented here should be applicable to the analysis of various target-enriched (e.g., amplicon or sequence capture) Illumina sequence data.

Genotyping by sequencing (GBS) provides opportunities to generate high-resolution genetic maps at a low genotyping cost, but for highly heterozygous species, missing data and heterozygote undercalling complicate the creation of GBS genetic maps. To overcome these issues, we developed a publicly available, modular approach called HetMappS, which functions independently of parental genotypes and corrects for genotyping errors associated with heterozygosity. For linkage group formation, HetMappS includes both a reference-guided synteny pipeline and a reference-independent de novo pipeline. The de novo pipeline can be utilized for under-characterized or high diversity families that lack an appropriate reference. We applied both HetMappS pipelines in five half-sib F₁ families involving genetically diverse Vitis spp. Starting with at least 116,466 putative SNPs per family, the HetMappS pipelines identified 10,440 to 17,267 phased pseudo-testcross (Pt) markers and generated high-confidence maps. Pt marker density exceeded crossover resolution in all cases; up to 5,560 non-redundant markers were used to generate parental maps ranging from 1,047 cM to 1,696 cM. The number of markers used was strongly correlated with family size in both de novo and synteny maps (r = 0.92 and 0.91, respectively). Comparisons between allele and tag frequencies suggested that many markers were in tandem repeats and mapped as single loci, while markers in regions of more than two repeats were removed during map curation. Both pipelines generated similar genetic maps, and genetic order was strongly correlated with the reference genome physical order in all cases. Independently created genetic maps from shared parents exhibited nearly identical results. Flower sex was mapped in three families and correctly localized to the known sex locus in all cases. The HetMappS pipeline could have wide application for genetic mapping in highly heterozygous species, and its modularity provides opportunities to adapt portions of the pipeline to other family types, genotyping technologies or applications.

Marker-assisted selection (MAS) is often employed in crop breeding programs to accelerate and enhance cultivar development, via selection during the juvenile phase and parental selection prior to crossing. Next-generation sequencing and its derivative technologies have been used for genome-wide molecular marker discovery. To bridge the gap between marker development and MAS implementation, this study developed a novel practical strategy with a semi-automated pipeline that incorporates trait-associated single nucleotide polymorphism marker discovery, low-cost genotyping through amplicon sequencing (AmpSeq) and decision making. The results document the development of a MAS package derived from genotyping-by-sequencing using three traits (flower sex, disease resistance and acylated anthocyanins) in grapevine breeding. The vast majority of sequence reads ( ⩾ 99%) were from the targeted regions. Across 380 individuals and up to 31 amplicons sequenced in each lane of MiSeq data, most amplicons (83 to 87%) had <10% missing data, and read depth had a median of 220–244×. Several strengths of the AmpSeq platform that make this approach of broad interest in diverse crop species include accuracy, flexibility, speed, high-throughput, low-cost and easily automated analysis.

Transferable DNA markers are essential for breeding and genetics. Grapevine (Vitis) breeders utilize disease resistance alleles from congeneric species ~20 million years divergent, but existing Vitis marker platforms have cross-species transfer rates as low as 2%. Here, we apply a marker strategy targeting the inferred Vitis core genome. Incorporating seven linked-read de novo assemblies and three existing assemblies, the Vitis collinear core genome is estimated to converge at 39.8 Mb (8.67% of the genome). Adding shotgun genome sequences from 40 accessions enables identification of conserved core PCR primer binding sites flanking polymorphic haplotypes with high information content. From these target regions, we develop 2,000 rhAmpSeq markers as a PCR multiplex and validate the panel in four biparental populations spanning the diversity of the Vitis genus, showing transferability increases to 91.9%. This marker development strategy should be widely applicable for genetic studies in many taxa, particularly those ~20 million years divergent.

Powdery mildews present specific challenges to phenotyping systems that are based on imaging. Having previously developed low-throughput, quantitative microscopy approaches for phenotyping resistance to Erysiphe necator on thousands of grape leaf disk samples for genetic analysis, here we developed automated imaging and analysis methods for E. necator severity on leaf disks. By pairing a 46-megapixel CMOS sensor camera, a long-working distance lens providing 3.5× magnification, X-Y sample positioning, and Z-axis focusing movement, the system captured 78% of the area of a 1-cm diameter leaf disk in 3 to 10 focus-stacked images within 13.5 to 26 seconds. Each image pixel represented 1.44 μm² of the leaf disk. A convolutional neural network (CNN) based on GoogLeNet determined the presence or absence of E. necator hyphae in approximately 800 subimages per leaf disk as an assessment of severity, with a training validation accuracy of 94.3%. For an independent image set the CNN was in agreement with human experts for 89.3% to 91.7% of subimages. This live-imaging approach was nondestructive, and a repeated measures time course of infection showed differentiation among susceptible, moderate, and resistant samples. Processing over one thousand samples per day with good accuracy, the system can assess host resistance, chemical or biological efficacy, or other phenotypic responses of grapevine to E. necator. In addition, new CNNs could be readily developed for phenotyping within diverse pathosystems or for diverse traits amenable to leaf disk assays.

Significance: We studied the grape sex-determining region (SDR) in 12Vitisgenomes and demonstrated its conservation across 556 genotypes including 193 accessions from 47 world-wide wild grapevine species and 363 accessions of cultivated grapevine. Although the grape SDR is recombination free in all wild species, we found two distinct hermaphrodite (H) haplotypes (H1 and H2) among the cultivated grapevines, both chimeras of male (M) and female (f) haplotypes. The two independent recombinations carry different genetic signatures which long predate the domestication of grapevine, suggesting independent evolutions of this trait in wild European grapevine gene pools prior to human domestication.