Two Novel er1 Alleles Conferring Powdery Mildew (Erysiphe pisi) Resistance Identified in a Worldwide Collection of Pea (Pisum sativum L.) Germplasms

Powdery mildew caused by Erysiphe pisi DC. severely affects pea crops worldwide. The use of resistant cultivars containing the er1 gene is the most effective way to control this disease. The objectives of this study were to reveal er1 alleles contained in 55 E. pisi-resistant pea germplasms and to develop the functional markers of novel alleles. Sequences of 10 homologous PsMLO1 cDNA clones from each germplasm accession were used to determine their er1 alleles. The frame shift mutations and various alternative splicing patterns were observed during transcription of the er1 gene. Two novel er1 alleles, er1-8 and er1-9, were discovered in the germplasm accessions G0004839 and G0004400, respectively, and four known er1 alleles were identified in 53 other accessions. One mutation in G0004839 was characterized by a 3-bp (GTG) deletion of the wild-type PsMLO1 cDNA, resulting in a missing valine at position 447 of the PsMLO1 protein sequence. Another mutation in G0004400 was caused by a 1-bp (T) deletion of the wild-type PsMLO1 cDNA sequence, resulting in a serine to leucine change of the PsMLO1 protein sequence. The er1-8 and er1-9 alleles were verified using resistance inheritance analysis and genetic mapping with respectively derived F2 and F2:3 populations. Finally, co-dominant functional markers specific to er1-8 and er1-9 were developed and validated in populations and pea germplasms. These results improve our understanding of E. pisi resistance in pea germplasms worldwide and provide powerful tools for marker-assisted selection in pea breeding.


Introduction
Pea (Pisum sativum L.) is a widely distributed legume crop, which frequently suffers from various stresses, including abiotic and biotic factors in the season of growth [1,2]. Powdery mildew, induced by Erysiphe pisi DC., severely reduces the yield and quality of pea crops worldwide [3][4][5]. Severe E. pisi infections of peas can lead to yield losses of up to 80% in regions which are suitable for disease development [5,6]. The use of resistant cultivars carrying the E. pisi-resistant gene er1 has been considered to be the most effective and environmentally friendly way to prevent this disease to date [6,7].
Formerly, E. pisi infection was the only known cause of pea powdery mildew. However, since 2005, two other Erysiphe species, Erysiphe trifolii and Erysiphe baeumleri, have been reported to also infect peas and induce the same powdery mildew symptoms as E. pisi in some regions [8][9][10]. Previous

Phenotypic Evaluation
Fifty-five E. pisi-immune or -resistant pea germplasm accessions from 13 countries were re-evaluated for their resistance to the E. pisi isolate EPYN. At 10 days post-inoculation, the E. pisi disease severity of all susceptible controls (Bawan 6 and Longwan 1) were rated as score 4. In contrast, the 55 E. pisi-resistant germplasm accessions appeared to be either immune (symptom-free; disease severity 0) or resistant (slight infection; disease severity 1-2) to E. pisi isolate EPYN. Of the 55 resistant germplasm accessions, 46 were classified as immune and nine as resistant to E. pisi (Table 1). To provide comprehensive information for the resistance of a worldwide collection of 86 pea germplasms to E. pisi, the phenotypes of 31 resistant pea germplasms carrying known er1 alleles are also shown in Table 1. "R", "I", and "S" stand for resistant, immune, and susceptible, respectively.
Two novel er1 alleles were discovered in the two remaining germplasms: G0004389 (from Afghanistan) and G0004400 (from Australia). A novel mutation pattern was found in the G0004389 cDNA fragment homologous to PsMLO1: a 3-bp deletion (GTG) corresponding to positions 1339-1341 in exon 15 (the final exon) of the PsMLO1 cDNA sequence. This deletion caused the loss of the amino acid valine at position 447 of the PsMLO1 protein sequence, probably resulting in a functional change ( Figure 1A). This mutation differed from all known er1 alleles, indicating that the E. pisi resistance of G0004389 was controlled by a novel allele of er1. This novel allele was designated er1-8, following the accepted nomenclature [14,26,27,42,44,51]. In pea germplasm G0004400, a 1-bp deletion (T) was identified in a previously unreported position homologous to position 928 in exon 10 of the PsMLO1 cDNA sequence. This deletion caused a substitution of the amino acid serine with leucine at position 310 of the PsMLO1 protein sequence ( Figure 1B). This change caused the early termination of protein translation, probably also resulting in a functional change of PsMLO1 ( Figure 1B). Thus, E. pisi resistance in G0004400 was also controlled by a novel er1 allele, herein designated er1-9. designated er1-8, following the accepted nomenclature [14,26,27,42,44,51]. In pea germplasm G0004400, a 1-bp deletion (T) was identified in a previously unreported position homologous to position 928 in exon 10 of the PsMLO1 cDNA sequence. This deletion caused a substitution of the amino acid serine with leucine at position 310 of the PsMLO1 protein sequence ( Figure 1B). This change caused the early termination of protein translation, probably also resulting in a functional change of PsMLO1 ( Figure 1B). Thus, E. pisi resistance in G0004400 was also controlled by a novel er1 allele, herein designated er1-9. Interestingly, frame shift mutations, where small fragments are deleted or inserted, were identified in the cloned sequences of several pea germplasms. The fragments homologous to the wild-type PsMLO1 cDNA in seven pea germplasms (G0002602, G0006515, G0002883, G0004448, G0002848, G0003935, and G0005117) had 5-bp deletions (GTTAG) at positions 700-704 of wild-type PsMLO1 cDNA, while three pea germplasm accessions (G0002883, G0002971, and L0368) had another 5-bp deletion (TAGGG) at positions 1235-1239 of the wild-type PsMLO1 cDNA. In accession G0006514, there was a 4-bp deletion (GGAG) at positions 181-184 of the wild-type PsMLO1 cDNA. In four pea accessions (G0002847, G0004434, G0003974, and Texuan 11) and two Interestingly, frame shift mutations, where small fragments are deleted or inserted, were identified in the cloned sequences of several pea germplasms. The fragments homologous to the wild-type PsMLO1 cDNA in seven pea germplasms (G0002602, G0006515, G0002883, G0004448, G0002848, G0003935, and G0005117) had 5-bp deletions (GTTAG) at positions 700-704 of wild-type PsMLO1 cDNA, while three pea germplasm accessions (G0002883, G0002971, and L0368) had another 5-bp deletion (TAGGG) at positions 1235-1239 of the wild-type PsMLO1 cDNA. In accession G0006514, there was a 4-bp deletion (GGAG) at positions 181-184 of the wild-type PsMLO1 cDNA. In four pea accessions (G0002847, G0004434, G0003974, and Texuan 11) and two pea accessions (G0002235 and G0002848), there were a 16-bp deletion (CTCATCTTCCTCCAGG) at positions 776-791 and a 16-bp insertion (AATTTTTCTGTTTCAG) at position 1171 of the wild-type PsMLO1 cDNA, respectively. In germplasm accession Jia 2, there was a 7-bp insertion (TAATAAG) at position 921 of the wild-type PsMLO1 cDNA. It was probable that these indels resulted from aberrant splicing events during transcription. Each frame shift mutation was observed in only one or two of ten cloned PsMLO1 cDNA sequences per germplasm accession.
Various alternative splicing patterns, including intron retention and exon skipping, were also observed in multiple PsMLO1 sequences cloned from the 55 resistant pea germplasm accessions. The eight introns retained were 1, 2, 4, 6, 7, 9, 12, and 13, and the three exons skipped were 4, 10, and 11 of the wild-type PsMLO1. Each intron retention and exon skipping event were discovered in only one or two of ten cloned PsMLO1 cDNA sequences.

Genetic Analysis and Mapping of er1-8 and er1-9
As expected, the two resistant pea parents, G0004389 and G0004400, were immune to E. pisi infection (disease severity 0), while the two susceptible parents (Bawan 6 and WSU 28) were heavily infected (disease severity 4) ( Figure 2). The segregation patterns of E. pisi resistance in the F 1 , F 2 , and F 2:3 populations derived from the crosses WSU 28 × G0004389 and Bawan 6 × G0004400 are presented in Table S1. pea accessions (G0002235 and G0002848), there were a 16-bp deletion (CTCATCTTCCTCCAGG) at positions 776-791 and a 16-bp insertion (AATTTTTCTGTTTCAG) at position 1171 of the wild-type PsMLO1 cDNA, respectively. In germplasm accession Jia 2, there was a 7-bp insertion (TAATAAG) at position 921 of the wild-type PsMLO1 cDNA. It was probable that these indels resulted from aberrant splicing events during transcription. Each frame shift mutation was observed in only one or two of ten cloned PsMLO1 cDNA sequences per germplasm accession. Various alternative splicing patterns, including intron retention and exon skipping, were also observed in multiple PsMLO1 sequences cloned from the 55 resistant pea germplasm accessions. The eight introns retained were 1, 2, 4, 6, 7, 9, 12, and 13, and the three exons skipped were 4, 10, and 11 of the wild-type PsMLO1. Each intron retention and exon skipping event were discovered in only one or two of ten cloned PsMLO1 cDNA sequences.
The cross of Bawan 6 × G0004400 generated five F1 plants which showed E. pisi-susceptibility (Table S1). One of five F1 plants generated 119 F2 offspring. 32 of 119 were resistant, and 87 of 119 were susceptible to E. pisi. The segregation ratio in the F2 population of resistance to susceptibility fitted a genetic model ratio of 1:3 (χ 2 = 0.14; P = 0.71), also indicating recessive heredity of a single Six F 1 plants produced from the cross WSU 28 × G0004389 were susceptible to E. pisi (Table S1). One of the six plants generated 120 F 2 and F 2:3 offspring through self-pollination. Of these 120 F 2 plants, 30 were resistant (R) to E. pisi, and 90 were susceptible (S) to E. pisi-. This indicates that the segregation ratio (resistance:susceptibility) in the F 2 population was exactly 1:3 (χ 2 = 0.01; P = 0.92), indicating recessive heredity of a single gene. Moreover, a segregation ratio of 30 (homozygous resistant): 63 (segregating): 27 (homozygous susceptible) in the F 2:3 population fitted well with the genetic model of 1:2:1 ratio (χ 2 = 0.48, P = 0.79) (Table S1), confirming that the E. pisi resistance in G0004389 was controlled by a single recessive gene.
The cross of Bawan 6 × G0004400 generated five F 1 plants which showed E. pisi-susceptibility (Table S1). One of five F 1 plants generated 119 F 2 offspring. 32 of 119 were resistant, and 87 of 119 were susceptible to E. pisi. The segregation ratio in the F 2 population of resistance to susceptibility fitted a genetic model ratio of 1:3 (χ 2 = 0.14; P = 0.71), also indicating recessive heredity of a single gene. Moreover, a segregation ratio of 32 (homozygous resistant): 64 (segregating): 23 (homozygous susceptible) in the F 2:3 population (119 families) fitted well with the genetic model of 1:2:1 ratio (χ 2 = 2.51; P = 0.29), indicating that E. pisi resistance in G0004400 was also controlled by a single recessive gene (Table S1).
Of the 20 markers tested, five (c5DNAmet, AD160, AA200, AA224, and PSMPSA5) were polymorphic between parents WSU 28 and G0004389, and seven (AC74, AD160, PSMPSAD51, ScOPD10-650, ScOPX04-880, ScOPE16-1600, and AD59) were polymorphic between Bawan 6 and G0004400, indicating that these markers were likely linked to the E. pisi resistance gene. Thus, the five and the seven parental polymorphic markers were used to confirm the genotypes of each F 2 plant derived from WSU 28 × G0004389 and Bawan 6 × G0004400, respectively. This genetic linkage analysis suggested that three markers (c5DNAmet, AA200, and AA224) and six markers (AD160, PSMPSAD51, ScOPD10-650, ScOPX04-880, ScOPE16-1600, and AD59) were linked to the resistance gene er1 in G0004389 and G0004400, respectively (Figure 3). Our results also indicated that the resistance genes in both germplasm accessions were located in the er1 region. In G0004389, the linkage map indicated that the markers (c5DNAmet and AA200) were mapped on both sides of the target gene with 9.6 cM and 3.5 cM genetic distances, respectively ( Figure 3A). In G0004400, two other markers (PSMPSAD51 and ScOPX04-880) were located on both sides of the target gene with 12.2 cM and 4.2 cM genetic distances, respectively ( Figure 3B). Our linkage and genetic map analyses confirmed that er1-8 and er1-9 controlled E. pisi resistance in G0004389 and G0004400, respectively ( Figure 3).  (Table S1). Of the 20 markers tested, five (c5DNAmet, AD160, AA200, AA224, and PSMPSA5) were polymorphic between parents WSU 28 and G0004389, and seven (AC74, AD160, PSMPSAD51, ScOPD10-650, ScOPX04-880, ScOPE16-1600, and AD59) were polymorphic between Bawan 6 and G0004400, indicating that these markers were likely linked to the E. pisi resistance gene. Thus, the five and the seven parental polymorphic markers were used to confirm the genotypes of each F2 plant derived from WSU 28 × G0004389 and Bawan 6 × G0004400, respectively. This genetic linkage analysis suggested that three markers (c5DNAmet, AA200, and AA224) and six markers (AD160, PSMPSAD51, ScOPD10-650, ScOPX04-880, ScOPE16-1600, and AD59) were linked to the resistance gene er1 in G0004389 and G0004400, respectively (Figure 3). Our results also indicated that the resistance genes in both germplasm accessions were located in the er1 region. In G0004389, the linkage map indicated that the markers (c5DNAmet and AA200) were mapped on both sides of the target gene with 9.6 cM and 3.5 cM genetic distances, respectively ( Figure 3A). In G0004400, two other markers (PSMPSAD51 and ScOPX04-880) were located on both sides of the target gene with 12.2 cM and 4.2 cM genetic distances, respectively ( Figure 3B). Our linkage and genetic map analyses confirmed that er1-8 and er1-9 controlled E. pisi resistance in G0004389 and G0004400, respectively ( Figure 3).

Validation and Application of Functional Markers
Of the 169 germplasm accessions selected and tested for their phenotypic resistance to E. pisi isolate EPYN (Table S2), 19 were phenotypically immune to E. pisi, 22 were resistant, and 128 were susceptible (Table S2).
To date, more than 40 MLO mutant alleles have been described in the monocotyledonous plant barley [52]. It is predicted that additional er1 alleles resulting from natural mutations would be present among pea germplasms from around the world. As expected, we not only encountered the four known er1 alleles (er1-1, er1-2, er1-6, and er1-7) across the 53 E. pisi-resistant pea germplasms, but we also discovered two novel er1 alleles: er1-8 in germplasm G0004389 from Afghanistan and er1-9 in germplasm G0004400 from Australia (Table 1).
Previous studies have indicated that the er1-2 allele produces three distinct PsMLO1 transcripts [14,25,27,51]. Interestingly, this study observed that the er1-2 carried by the pea germplasm accession G0002860 produced four distinct PsMLO1 transcripts. One of these transcripts was characterized by a 129-bp deletion, corresponding to the deletion of exon 13 (68 bp) and exon 14 (61 bp) from wild-type PsMLO1 cDNA, indicating alternative splicing of exon skipping. Previously, two transcripts of er1-2 were observed to have large insertions (155-bp and 220-bp) based on comparisons with the transcripts of wild-type PsMLO1 cDNA [14,24,25,27,51]. Here, we discovered that the 155-bp "insertion" in er1-2 resulted from a 192-bp insertion at position 1263 and a 37-bp deletion of positions 1263-1299 in exon 14 of wild-type PsMLO1, while the 220-bp "insertion" resulted from a 257-bp insertion at position 1263 and a 37-bp deletion of positions 1263-1299 in exon 14 of wild-type PsMLO1. Another alternative transcript of er1-2, an 87-bp "insertion", was observed and resulted from a 192-bp insertion and a 37-bp deletion in exon 14 and a 68-bp deletion corresponding to exon 13 of wild-type PsMLO1. Our blast analysis indicated that the 192-and 257-bp insertions had 95% sequence identity with a five-part repetition in the pea genomic BAC sequence (GenBank accession number CU655882). These insertions were also highly similar (~85-87% identity) to a portion of the giant Ogre retrotransposons in the pea genome (GenBank accession numbers AY299395, AY299398, AY299397, and AY299394).
Alternative splicing in eukaryotes is a pervasive molecular mechanism that significantly increases transcriptome and proteome complexity [53]. Four main types of alternative splicing are known: exon skipping, alternative 5 splice sites, alternative 3 splice sites, and intron retention [54]. Exon skipping is common in humans, while intron retention is common in plants [55]. Alternative splicing is involved in many physiological processes, including response to biotic and abiotic stressors [56]. In the pea germplasms, three types of alternative splicing, intron retention, exon skipping, and alternative 5 splice site selection, were observed in this study. Interestingly, pea germplasms carrying identical er1 alleles varied in their resistance to E. pisi, from immune (disease severity of 0) to merely resistant (disease severity of 1-2) ( Table 1). Alternative splicing in response to biotic stress may affect the expression of regulatory genes. Thus, it is speculated that the alternative splicing of er1 alleles might affect the expression of the E. pisi resistance genes er1. In addition, the different levels of resistance to E. pisi might result from other related gene regulation. It is possible that multiple molecular processes and pathways contribute to MLo-based E. pisi resistance in peas.
This study discovered two novel er1 alleles resulting from novel mutations of wild-type PsMLO1 cDNA: er1-8 was generated by a 3-bp deletion in exon 15, and er1-9 was generated by a 1-bp deletion in exon 10. The co-dominant functional markers specific to er1-8 (InDel-er1-8 and KASPar-er1-8) and to er1-9 (KASPar-er1-9) were developed. These markers were validated in genetic populations and in pea germplasms. Our results are vital for future studies of powdery mildew resistance and for the development of E. pisi-resistant pea cultivars. The novel er1 alleles and the corresponding co-dominant functional markers developed herein could constitute efficient and powerful tools for the breeding of E. pisi-resistant peas.
The E. pisi isolate EPYN from Yunnan Province of China was used as the inoculum [26,27,41,48,50,51]. The EPYN isolate was maintained through continuous re-inoculation of seedlings of the pea cultivar Longwan 1 under controlled conditions. The inoculated plants were incubated in a growth chamber to prevent contamination with other isolates [25].

Phenotypic Evaluation
Twenty seeds were planted from each of the 55 E. pisi-resistant pea germplasm accessions, from the susceptible controls Bawan 6 and Longwan 1, and from the resistant controls Xucai 1 and YI [27]. The seedlings were thinned to 15 per pot before the phenotypic evaluation. Three replications were planted. Seeded pots were placed in a greenhouse maintained at 18 to 26 • C. At the same time, the E. pisi inoculum was prepared by inoculating the 10-day-old seedlings of the susceptible pea cultivar Longwan 1, which were incubated in a growth chamber at 20 ± 1 • C with a 12-h photoperiod. Two weeks later, the 14-day-old seedlings of 55 germplasm accessions and controls were inoculated by gently shaking off conidia of the Longwan 1 plants. Inoculated plants were incubated in a growth chamber at 20 ± 1 • C with a 12-h photoperiod. Ten days later, disease severity was rated based on a scale (0-4 scale) [27]. Plants with a score of 0 were considered E. pisi-immune, while those with scores of 1 and 2, 3 and 4 were considered as E. pisi-resistant and E. pisi-susceptible, respectively. For those identified as immune or resistant to E. pisi, repeated identification was performed.

RNA Extraction and PsMLO1 Sequence Analysis
The extraction of total RNA and synthesis of cDNA from the 55 pea germplasms and controls were completed according to our previous studies [25][26][27].
To identify the resistance alleles at the er1 loci, the full-length cDNAs of the PsMLO1 homologs were amplified using the primers specific for PsMLO1 [14]. The PCR cycling conditions were as follows: 95 • C for 5 min; then 35 cycles of denaturation at 94 • C for 30 s, annealing at 58 • C for 45 s, and extension at 72 • C for 1 min; and a final extension at 72 • C for 10 min. The purified amplicons were cloned with a pEasy-T5 vector (TransGen Biotech, Beijing, China). The sequencing reactions of 10 clones per germplasm (including controls) were performed by the Shanghai Shenggong Biological Engineering Co., Ltd. (Shanghai, China). The resulting sequences were aligned with wild-type PsMLO1 of pea (NCBI accession number: FJ463618.1) using DNAMAN v6.0 (Lynnon Biosoft, Quebec, Canada).

Genetic Analysis of Pea Germplasms Carrying Novel Alleles
To confirm the resistance genes, er1-8 and er1-9, G0004389 and G0004400 were crossed with the E. pisi-susceptible cultivars WSU 28 and Bawan 6, respectively, to generate genetic populations. The derived F 1 , F 2 , and F 2:3 populations from both crosses (WSU 28 × G0004389 and Bawan 6 × G0004400) were used to evaluate the E. pisi resistance and genetic analysis of G0004389 and G0004400. The four parents and the derived F 1 and F 2 populations were planted in a propagation greenhouse to generate F 2 and F 2:3 family seeds, respectively.
Plants of the F2 populations at the fourth or fifth leaf stage were inoculated with the E. pisi isolate EPYN using the detached leaf method [25][26][27]57]. After inoculation, the treated leaves were placed in a growth chamber at 20 • C with a 14-h photoperiod. The four parents (WSU 28, G0004389, Bawan 6, and G0004400) were also inoculated as controls. Ten days after inoculation, disease severity was rated based on a scale of 0-4 as described above. Plants with scores of 0-2 and 3-4 were classified as resistant and susceptible, respectively [25][26][27]31,58]. Those plants identified as E. pisi-resistant were tested again to confirm their resistance.
Twenty-five seeds were selected randomly from each of the 120 F 2:3 families derived from WSU 28 × G0004389, and from each of the 119 F 2:3 families derived from Bawan 6 × G0004400. These seeds were planted and cultivated together with their parents, following previously published protocols [25][26][27]. Disease severity was scored 10 days after inoculation using the 0-4 scale, as described above for the phenotypic identification of the pea germplasms. The F 2:3 families with scores of 0-2 and 3-4 were classified as homozygous resistant and homozygous susceptible, respectively. Families with scores of 0-2 and 3-4 were considered segregated to E. pisi resistance [27,31,58]. The families identified as homozygous resistant or resistance segregated were subjected to repeated testing.
A chi-squared (χ 2 ) analysis was used to evaluate the goodness-of-fit to Mendelian segregation ratio of the F 2 and F 2:3 phenotypes derived from WSU 28 × G0004389 and Bawan 6 × G0004400.

Genetic Mapping of the Resistance Alleles er1-8 and er1-9
The Genomic DNA was isolated from the leaves of the F 2 populations and of their parents using the cetyltrimethylammonium bromide (CTAB) extraction method [59]. The DNA solution was diluted and stored at −20 • C until use.
The segregation data of the polymorphic markers in the F 2 populations were evaluated for goodness-of-fit to Mendelian segregation patterns with a chi-squared (χ 2 ) test. Genetic linkage analyses were completed using MAPMAKER/EXP version 3.0b. A logarithm of odds (LOD) score > 3.0 and a distance < 50 cM were used as the thresholds to determine the linkage groups [62]. Genetic distances were determined using the Kosambi mapping function [63]. The genetic linkage map was constructed using the Microsoft Excel macro MapDraw [64].

Development of Functional Markers for er1-8 and er1-9
Primers flanking the mutation site (GTG/-) were designed based on the PsMLO1 gene sequence (GenBank accession number KC466597), using Primer Premier v5.0, to develop an insertion/deletion (indel) functional marker specific to allele er1-8, InDel-er1-8 ( Table 3). The marker InDel-er1-8 was used to determine the genotypes of the 120 F 2 offspring derived from WSU 28 × G0004389. PCR amplification was performed as described above on a thermal cycler with the following cycling program: 95 • C for 5 min; 35 cycles of 94 • C for 30 s, 55 • C for 30 s, and 72 • C for 30 s; and 72 • C for 7 min. PCR products were separated on 8% polyacrylamide gels. Table 3. Sequence information for the indel and Kompetitive allele-specific PCR (KASPar) markers specific to er1-8, and for the KASPar marker specific to er1-9.
KASPar markers were amplified with a Douglas Scientific Array Tape Platform (China Golden Marker, Beijing, Biotech Co., Ltd.) in a 0.8 µL Array Tape reaction volume with 10 ng dry DNA, 0.8 µL 2 × KASP master mix, and 0.011 µL primer mix (KBioscience, Hoddesdon, UK). A Nexar Liquid handling instrument was used to add the PCR solution to the Array Tape (Douglas Scientific). PCRs were performed on a Soellex PCR Thermal Cycler with the following conditions: initial denaturation at 94 • C for 15 min; followed by 10 cycles of denaturation at 94 • C for 20 s, and 65 • C for 60 s at an annealing temperature that decreased by 0.8 • C per cycle; and then 26 cycles of denaturation at 94 • C for 20 s and 57 • C for 60 s; and a final cooling to 4 • C. A fluorescent end-point reading was completed with the Araya fluorescence detection system (part of the Douglas Scientific Array Tape Platform). Genotypes and clusters were visualized with Kraken (http://ccb.jhu.edu/software/kraken/MANUAL.html).