Initial resistance screening was conducted between March 20th and April 8th, 2019, in a greenhouse in Athens, GA. Seed from the parents, F1 and 111 F2:3 families were sown in 48-well seedling trays in a randomized complete block design with four plants per block and four blocks. Two additional trays containing controls and 12 of the F2:3 families were also sown for mock inoculations. Seedlings were grown in a greenhouse with supplemental lighting until the 2–3 true leaf stage (16 days) and inoculated by spraying inoculum on the seedlings with a hand spray bottle until run-off. The mock inoculations were sprayed with 0.1% Tween-20 solution. After inoculation, the plants were placed inside a plastic tunnel with two humidifiers (Trion IAQ, Sanford, NC) in the greenhouse (average temperature = 24.0°C, average humidity = 98.8%). After 72 hours, the seedlings were moved to a greenhouse bench and overhead watered twice per day for four days. At 7 days post inoculation, the seedlings were rated for disease severity using a 0–10 scale (0: no lesions present on the first two true leaves, 1: 1–10% of first two true leaves covered in lesions, 2: 11–20% of first two true leaves covered in lesions, 3: 21–30% of first two true leaves covered in lesions, 4: 31–40% of first two true leaves covered in lesions, one leaf beginning to collapse, 5: 40–50% of first two true leaves covered in lesions, 1 true leaf collapsed, 6: 1 of first two true leaves dead, other 1–25% of true leaves covered in lesions, 7: 1 of first two true leaves dead, other 26–50% covered in lesions, 8: 1 of first two true leaves dead, other 51–75% covered in lesions, 9: both first two true leaves dead, 10: seedling totally collapsed or dead). The mean severity rating for each family/control was calculated using JMP® version 15.0.0 (SAS Institute Inc., Cary, NC). A Shapiro-Wilk test was conducted using JMP to test the normality of the phenotypic distribution. The 20 most resistant and 20 most susceptible families were re-screened using the same protocol between May 1st and May 25th, 2019 (average temperature = 23.4°C, average humidity = 91.7%). The 12 most resistant and most susceptible families over both screens were used to construct the bulks for QTL-seq. All methods were performed in accordance with the institutional and national Biosafety regulations.
DNA Extraction and Sequencing.
Frozen leaf material from the F2 plants of the 24 selected F2:3 families was ground using a TissueLyser II (QIAGEN, Hilden, Germany) and DNA was extracted using the E.Z.N.A. HP Plant DNA Mini Kit (Omega Bio-Tek Inc., Norcross, GA). DNA was quantified with an Infinite M200 Pro plate reader (Tecan Group Ltd., Mannedorf, Switzerland) using the i-control software (Tecan Group Ltd.,) and a NanoQuant Plate™ (Tecan Group Ltd.,). Agarose gel electrophoresis was used to confirm the quality of genomic DNA. Equal amounts of DNA from the 12 most resistant and 12 most susceptible families were pooled to create the resistant (R-Bulk) and susceptible bulk (S-Bulk), respectively. The samples were sent to Novogene (Novogene Corporation Inc., Davis, CA) for whole genome sequencing on an Illumina Platform (PE150, Q30 > 80%).
The raw reads were combined into forward and reverse read files for each bulk and checked for quality using FastQC41. The reads were then aligned to the C. lanatus 97103_v2 genome using BWA-MEM24,42. The resulting SAM files were converted into BAM files using SAMtools, which was also used to sort the reads by alignment position and then index the resulting BAM files43. SAMtools was used to calculate mapping and pairing ratios of raw reads. BEDtools was then used to calculate average read depth for the BAM files44. Read group names were standardized using Picard Tools (http://broadinstitute.github.io/picard/), which was then used to mark duplicate reads and index the files. SAMtools was used to index the 97103_v2 genome, and Picard Tools was used to create a dictionary file for the 97103_v2 genome. Reads were re-aligned using GATK to generate clean reads from misaligned regions45. GATK was used to perform variant calling and the resulting VCF files from each bulk were combined into a single VCF file. GATK was used to filter the VCF file so that only SNPs remained. These SNPs were then filtered using GATK (QD < 2.0 || FS > 60.0 || MQ < 40.0 || MQRankSum < -12.5 || ReadPosRankSum < -8.0), and filtered SNPs were then removed using VCFtools46. GATK was used to output the SNPs from the VCF file into a table format suitable for use in the R package QTLseqr47,48. After importation into QTLseqr, SNPs were again filtered (refAlleleFreq = 0.2, minTotalDepth = 100, maxTotalDepth = 150, minSampleDepth = 40, minGQ = 99), and then used to calculate the Δ-SNP index at each SNP19. This Δ-SNP index was used with a 1Mb sliding window to calculate a smoothed Δ-SNP index as well as 95% and 99% confidence intervals for a region’s contribution to the trait of interest (GSB resistance).
Primer Design and Genotyping.
KASP (LGC Genomics LLC, Teddington, UK) primers were designed for SNPs spanning the regions of interest identified by QTLseqr using Primer3Plus and the C. lanatus 97103_v2 genome24. DNA was extracted for parents, F1, and all F2 plants as previously described. KASP PCR reactions contained 1.94 µL of 2 × KASP Master Mix (LGC Genomics LLC), 0.06 µL of KASP Primer Mix, and 2 µL of DNA (10–20 ng/µl) in a total volume of 4 µL. The KASP Primer Mix contained 12 µL of each forward primer (100 µM), 30 µL of reverse primer (100 µM), and 46 µL of sterile distilled water. The following PCR conditions were used: 95°C for 15 minutes followed by 10 touchdown cycles of 95°C for 20 seconds, 66°C for 25 seconds, and 72°C for 15 seconds, followed by 35 cycles of 95°C for 10 seconds, 57°C for 60 seconds and 72°C for 15 seconds. KASP fluorescent end readings were measured using an Infinite M200 Pro (Tecan Group Ltd.) plate reader using the Magellan (Tecan Group Ltd.) software. Genotypes were called using KlusterCaller™ (LGC Genomics LLC).
A genetic map (n = 111) was created of each region of interest using the following settings in ICIMapping49: Grouping was performed by recombination frequency with a threshold value of 0.30, ordering was performed k-Optimally by recombination frequency using the 2-OptMAP algorithm with 10 NN initials, rippling was performed by recombination frequency with a window size of 5. Resulting genetic maps were visualized using MapChart50. The genetic map was then used for QTL mapping using ICIMapping with the following settings49: The ICIM-ADD mapping method was used with deletion of missing phenotypes, a 1cM step, and a value of 0.001 for the probability in stepwise regression. The LOD threshold of 2.5079 was determined by running 1000 permutations. The resulting data was graphed using R and the ggplot package48,51. A Tukey-Kramer test was performed using JMP ® version 15.0.0 (SAS Institute Inc., Cary, NC) to test the association between flanking markers, as well as a haplotype representing both markers simultaneously, and GSB disease severity in the population.