Genotyping-by-Sequencing in Vitis

Methods from tissue sampling through SNP assignment were as described previously (Hyma et al. 2015 ). For each vine, a single, small, newly expanding leaf (less than 1-cm diameter) was collected in one tube of a Costar 96-well cluster tube collection plate (Corning Life Sciences, Tewksbury, MA, USA). Leaf tissues were maintained at 4 °C from harvest until processing in the laboratory where two stainless steel genogrinder beads were placed in each tube and plates were frozen at −80 °C. Tissue grinding took place in a Geno/Grinder 2000 (OPS Diagnostics LLC, Lebanon, NJ, USA) with 96-well plates agitated in pairs at 400× speed for 1 min. Plates were then stored at −80 °C until processing with DNeasy 96-well DNA extraction kits (QIAGEN, Valencia, CA, USA). Modifications were made to the manufacturer’s protocol to improve DNA quality and quantity as follows: (1) PVP-40 (2 % w/v) was added to the AP1 lysis buffer prior to heating and (2) visual inspection for complete re-suspension of the sample pellet of each 8-tube strip was added to the agitation step following AP1 addition.
GBS was performed as described by Elshire et al. (2011 (link)), integrating four 96-well plates across 384 bar codes for library preparation (Hyma et al. 2015 ) and sequencing of single-end 100-bp sequences using a HiSeq 2000 (Illumina Inc., San Diego, CA, USA). Raw data for the University of Minnesota germplasm (henceforth, UMN dataset) was processed based on an updated version of the VitisGen databased previously employed by Hyma et al. (2015 ). Comprised of 409 maternal half-sib individuals and four progenitors, the UMN dataset was processed through TASSEL 3.0.139 GBS pipeline (Glaubitz et al. 2014 ) using the 12×.2 V. vinifera “PN40024” reference genome (Adam-Blondon et al. 2011 ; Jaillon et al. 2007 (link)) from The French-Italian Public Consortium (https://urgi.versailles.inra.fr/Species/Vitis/Data-Sequences/Genome-sequences). For alignment, Burrows–Wheeler Aligner maximal exact match (BWA-MEM) was applied using with default parameters (Li and Durbin 2009 ). The output consisted of variant call format (VCF) file version 4.1 (Danecek et al. 2011 ) including SNPs present in at least 40 % of the progeny, and with a minor allele frequency (MAF) ≥0.1. Subsequently, the VCF file was filtered using vcftools ver. 1.12a (Danecek et al. 2011 ) and TASSEL (Bradbury et al. 2007 ) versions 3.0.139 and 4.3.13. Finally, the VCF file was sliced to include SNP data of the 277 F₁ individuals (in mapping families of GE0711/1009 and GE1025) and the three progenitors (MN1264, MN1214, and MN1246) in this study.