Content uploaded by Giulia Pagliarani
Author content
All content in this area was uploaded by Giulia Pagliarani on Jun 03, 2016
Content may be subject to copyright.
ORIGINAL ARTICLE
Development of SNP-based assays for disease resistance and fruit
quality traits in apple (Malus ×domestica Borkh.) and validation
in breeding pilot studies
Isabelle O. Baumgartner
1
&Markus Kellerhals
1
&Fabrizio Costa
2
&Luca Dondini
3
&
Giulia Pagliarani
3
&Roberto Gregori
3
&Stefano Tartarini
3
&Lucie Leumann
1
&
François Laurens
4
&Andrea Patocchi
1
Received: 23 November 2015 /Revised: 21 January 2016 /Accepted: 22 March 2016
#Springer-Verlag Berlin Heidelberg 2016
Abstract The development of molecular markers linked to
specific traits is now routine practice, but the gap between
genomics and breeding often delays their application. In the
frame of the FP7 European project FruitBreedomics, apple
pilot studies were designed to exploit the project’soutcomes
towards the practical application of marker-assisted breeding
(MAB) programs. The aim of this pilot study was to develop
an outsourcing genotyping pipeline, which will provide access
to the single nucleotide polymorphism (SNP) markers analy-
sis for breeding companies without an internal DNA lab. The
process from seed sowing to genotypic and phenotypic seed-
ling selection was optimized. KASP™(competitive allele-
specific PCR) genotyping assays were developed for a num-
ber of major resistance genes for apple scab (Rvi2,Rvi4,Rvi6,
and Rvi15); powdery mildew (Pl2); and rosy apple aphid (Dp-
fl). In addition, KASP™assays for the genes Md-ACS1,Md-
ACO1,andMd-PG1 involved in fruit quality (firmness, tex-
ture, and storability) were also developed. The pilot study
demonstrated the efficacy of the SNP-based selection strategy,
especially for those programs dealing with traits not easily
assessable in vivo, such as pyramided resistances and fruit
quality traits.
Keywords MAB .Gene pyramiding .Venturia inaequalis .
Podosphaera leucotricha .Dysaphis plantaginea .
Md-ACS1 .Md-ACO1 .Md-PG1
Introduction
Most apple breeding programs worldwide aim to select novel
accessions distinguished by excellent eating quality, high
yield, and durable disease resistance. To achieve these goals,
efficient selection methods are crucial. Until recently, apple
breeding relied mainly on phenotypic selection and screening
methods such as phenotypic screening for disease and pest
resistance, tree and yield characteristics, and fruit quality.
Marker-assisted breeding (MAB) is an efficient strategy
especially for the identification of genotypes with multiple
resistance genes (R-genes) and for the early selection of
fruit-related traits that cannot be assessed at the seedling stage.
Despite the number of scientific reports describing marker-
trait association, their actual use in breeding programs remains
limited for different crops (Collard and Mackill 2008). The
FP7 European project FruitBreedomics (Laurens et al. 2010;
www.fruitbreedomics.com) aimed to bridge the gap between
genomics and breeding. The major goal was to modernize
breeding programs, therefore increasing the success rate of
accessions developed with the application of MAB. An
initial survey was conducted to identify the priority breeding
objectives among different European apple breeders, in order
to develop a common breeding strategy. Pilot studies were
then initiated, taking into account biotic and abiotic traits,
Communicated by E. Dirlewanger
*Andrea Patocchi
andrea.patocchi@agroscope.admin.ch
1
Agroscope Research Station, Plant Protection and Extension Fruits
and Vegetables, Schloss, P.O. Box 8820, Waedenswil, Switzerland
2
Research and Innovation Centre, Fondazione Edmund Mach, Via
Mach 1, 38010 San Michele all’Adige, Trento, Italy
3
Department of Agricultural Sciences, University of Bologna, Viale
Giuseppe Fanin 46, 40127 Bologna, Italy
4
INRA Angers, Institut de Recherche en Horticulture et Semences
UMR1345 (INRA/Agrocampus-ouest/Université d’Angers), 42 rue
Georges Morel-BP 60057, 49071 Beaucouzé cedex, France
Tree Genetics & Genomes (2016) 12:35
DOI 10.1007/s11295-016-0994-y
and the availability of both plant materials and markers linked
to specific traits.
The project FruitBreedomics has already released a com-
prehensive list of single nucleotide polymorphism (SNP)
markers associated with apple scab, powdery mildew, rosy
apple aphid (RAA), and fire blight resistance loci to the sci-
entific community (Jänsch et al. 2015; Padmarasu et al. 2014;
Cova et al. 2015;Pagliaranietal.2016). In addition, the US
Project BRosBREED^has also released SNP markers linked
to apple traits. Guan et al. (2015), for instance, identified sev-
eral quantitative trait loci (QTL) and SNP markers associated
with sugars and soluble solids content in apple. Currently, the
number of apple SNP markers is constantly increasing be-
cause of the available apple genome sequence and the reduc-
ing costs of SNP genotyping (Chagné et al. 2012; Bianco et al.
2014). MAB in apple will evolve from selection mainly based
on sequence characterized amplified region (SCAR) and sim-
ple sequence repeat (SSR) markers to SNPs. Although SCAR
and SSR markers are easy to use for the screening of seedlings
and germplasm collections, they are difficult to assemble in
large multiplex sets, since usually no more than six markers
can be combined in the same PCR reaction. Furthermore, their
analyses are not automatable in modern high-throughput
genotyping systems. In contrast, SNP markers are very flexi-
ble. Depending on the purpose of the work, they can be
exploited on platforms (e.g., Illumina, Affimetrix) suited to
the simultaneous assessment of thousands of SNP markers.
Alternatively, other methods, such as KASP™(competitive
allele-specific PCR) technology, Fluidigm SNP type Assay
and Sequenom MassARRAY Analyzer (Thomson 2014;
Varshney et al. 2009), enable the analysis of a few SNPs on
large progenies.
Currently, MAB in apple is mainly applied for the identifi-
cation of seedlings carrying a pyramided set of resistance genes
(e.g., apple scab, powdery mildew, and fire blight resistance
loci; Bus et al. 2009; Baumgartner et al. 2015) and improved
fruit quality traits (Zhu and Barritt 2008;Longhietal.2013b)
using SSR or SCAR markers. For apple scab (Rvi2,Rvi4,Rvi5,
Rvi6,Rvi11,Rvi12,andRvi15); powdery mildew (Pl2); fire
blight (FB_E,FB_MR5); and RAA (Dp-fl) resistance genes,
linked SNP markers have already been found (Jänsch et al.
2015; Padmarasu et al. 2014; Pagliarani et al. 2016; Cova
et al. 2015). For fruit quality, with the exception of the work
of Guan et al. (2015), to our knowledge, no SNP have been
reported. To this end, the majority of the efforts were dedicated
to the development of markers linked to firmness and ethylene
production, two physiologically aspects tightly correlated in
apple, and playing a crucial role for the postharvest perfor-
mance as well as for the general consumer’s appreciation
(Costa et al. 2011,2012). For fruit firmness, a marker was
developed for Md-PG1, a polygalacturonase gene involved
in the cell wall disassembly process. The allelic variability of
this marker is related to a SSR motif, leading to the
characterization of three alleles (Longhi et al. 2013a,b). For
ethylene, a hormone triggering several fruit ripening physio-
logical processes, in particular fruit softening, two functional
markers have been published to date, namely Md-ACS1
(Costa et al. 2005;Haradaetal.2000)andMd-ACO1(Costa
et al. 2005; Zhu and Barritt 2008), whose alleles depend on an
insertion or deletion mutation (INDEL) type of polymorphism.
Allelism of these two markers has been correlated with the
production of ethylene, and it has been validated and tested
in several breeding activities (Longhi et al. 2013b;Zhuand
Barritt 2008). The opportunity to pre-select for low ethylene
production and high firmness maintenance (with reduced deg-
radation of the cell wall) would improve fruit storability, lead-
ing to improved overall quality.
An aim of this pilot study was to develop a protocol
allowing any breeder, including those without their own
DNA laboratory, to implement MAB in their breeding pro-
grams. To facilitate this, all steps including parental selection,
seed sowing, and sampling of plant material for molecular
analysis, shipment of the samples to the genotyping company,
and finally, the identification of the seedlings matching the
desired ideotype according to the marker prediction, were op-
timized and tested. This work describes the initially required
development and validation of SNP assays for the apple scab
resistance genes Rvi2,Rvi4,Rvi6, and Rvi15; the powdery
mildew gene Pl2; the rosy apple aphid (RAA) resistance
Dp-fl;andforthethreegenesMd-ACS1,Md-ACO1,and
Md-PG1 involved in ethylene production and fruit softening.
The results from the genotyping and phenotyping of large
pilot study populations for the selected traits as well as sug-
gestions for the optimization of marker-assisted seedling se-
lection are presented.
Material and methods
Pilot study plant material
For the pilot study, 10 progenies were established and made
available at Agroscope (Wädenswil, Switzerland) and
Bologna University (Department of Agricultural Sciences,
Italy). Resistance sources found in parents were validated
and traced back to the progenitors by pedigree analyses
(Pedimap software, Voorrips et al. 2012.).
At Agroscope, cross A between ACW 11303 ((BArlet^×
BGloster^)×BRewena;^Rvi6)andACW18522(BTop az^×
(BMilwa^×BReka^); Rvi2 and Rvi6) and cross B between
ACW 13652 (BLa Flamboyante^× (A163-42 × A810-222);
Rvi6 and Pl2) and ACW 11567 (Milwa × Reka; Rvi2)were
made in spring 2011. Seeds were harvested from mature fruit
in autumn, and subsequently frozen at −20 °C. In December
2012, the seeds were stratified in humid sand at 2 °C for
2 months and in February 2013, the seeds of these crosses
35 Page 2 of 21 Tree Genetics & Genomes (2016) 12:35
were sown in 4 × 6 well trays (Quick Pot QP 24 T/13,
HerkuPlast, Kubern GmbH, Germany) and grown in a green-
house. Four 4 × 6 well trays in a block align to the 8 × 12
format of 96-well plates used for leaf sampling and molecular
analysis. Seedlings (2087 and 2578 for cross A and B, respec-
tively) were identified by the number given to the virtual 96-
well plate formed by the four trays and their position (1 to 12
and A to H) within the 96-well plates.
At the University of Bologna, eight fruiting progenies (C to
J) were selected from the ongoing breeding program. In the
first set of four progenies, BGoldRush^(Rvi6 and Dp-fl)was
always used as female parent and crossed with: BRealka^
(Rvi2 and Rvi4; cross C, 136 seedlings), BDiscovery^(scab
QTL on LG17]; cross D, 190 seedlings), GMAL2473 (Rvi15;
cross H, 95 seedlings), and BWhite Angel^(Plw); cross J, 93
seedlings). In the second group, instead, BFlorina^(also Rvi6
and Dp-fl) was used as female parent and crossed with BFuji^
(cross E, 86 seedlings), BGrenoble^(cross F, 56 seedlings),
BRoyal Gala^(cross G, 160 seedlings), and BPerleberg^(cross
I, 207 seedlings). All the seedlings were grafted on M.9
rootstock.
Selection strategy applied in the pilot studies
According to the segregation of the apple scab resistance
genes in progenies A and B, 12.5 and 25 % of the seedlings,
respectively, were expected to be susceptible, thus not carry-
ing any apple scab resistance gene. To reduce the number of
seedlings to be genotyped with SNP markers, the seedlings
were first inoculated with scab at the four-leaf stage as de-
scribed by Baumgartner et al. (2015). The conidia used for
inoculum production were collected from infected leaves of
the 2012 glasshouse seedling screening, which originated
from infected leaves collected from susceptible trees in or-
chards from different regions of Switzerland that were not
treated with fungicides. The concentration of the suspension
wasadjustedto4.3×10
5
conidia ml
−1
(germination rate
82 %). Fourteen days after inoculation (DAI), all seedlings,
with the exception of those kept as controls, clearly showing
class 4 symptoms (Chevalier et al. 1991) were removed from
the trays. The resultant Bholes^in the trays were filled with
resistant seedlings in order to minimize the number of 4 × 6
well plates. After SNP genotyping, the seedlings of progenies
A and B carrying the pyramided apple scab resistance genes
Rvi2 and Rvi6 (independently if Rvi6 was at the homo- or
heterozygous state in cross A) and the powdery mildew resis-
tance gene Pl2 (for cross B only) were selected. For the vali-
dation of the KASP
TM
assays for Rvi2 and Rvi6, a subset of
seedlings carrying only Rvi2,Rvi6, or neither resistance gene
(susceptible genotypes) from both crosses were kept as con-
trols. The seedlings, transplanted in 3-l pots, were then grown
till late summer without fungicide treatments. A phenotypic
selection was then performed considering (low) severity of
powdery mildew, a minimum vigor of the seedlings, and
avoiding plants with strong juvenility symptoms such as small
leaf size and spines.
Selected plants of progenies A and B were grafted onto
rootstock M.27 with a BSchneiderapfel^interstem and grown
for one season in the tree nursery and finally planted in an
orchard.
Regarding the orchard-grown progenies from Bologna (C
to J), most of the seedlings had previously gone through sev-
eral steps of scab resistance selection either in the greenhouse
(crosses C and D) or orchard (crosses E, F, G, and I). The
greenhouse scab resistance selection of crosses C and D was
made with a standard inoculation procedure (Paris et al. 2009)
and the seedlings scored as 2/3a according to Chevalier et al.
(1991) were grafted and planted in an orchard. For these two
crosses, the leaves from all plants in the field were collected
and included in the study. The field evaluation was made by
scoring the leaf sporulation on each plant grown in an un-
sprayed orchard. In order to reduce the number of plants to
be genotyped, for crosses E, F, G, and I, only leaves from the
resistant plants (no symptoms) were collected with the excep-
tion of a few susceptible controls from each population nec-
essary for the validation of the KASP
TM
assays. The H
(GoldRush × GMAL2473) and J (GoldRush × White Angel)
progenies were not pre-selected for scab resistance so leaves
from all plants were collected and included in the study, inde-
pendently of the level of resistance/susceptibility. Therefore,
25 and 50 % of the seedlings for crosses H and J, respectively,
were expected to be susceptible (not carrying any scab resis-
tance gene).
Identification and validation of SNP markers
for Md-ACO1,Md-ACS1,andMd-PG1
While developing the apple 20K FruitBreedomics SNP
Illumina array (Bianco et al. 2014), candidate SNPs associated
with Md-ACS1 (MDP0000370791, SNPs: FEM_cg_5 to 7);
Md-ACO1 (MDP0000195885; SNPs: FEM_cg_1 to 4); and
Md-PG1 (MDP0000326734; SNPs: FEM_cg_8 to 21) were
included. SNPs related to these genes were identified by full-
length re-sequencing in the apple cultivars BGolden
Delicious^,BDelearly^,BGranny Smith^,BBraeburn^,
BCripps Pink^,BScarlet^,BRoyal Gala^,andBFuji^.
Only SNPs flanked by a region of 60 bp, both up and
downstream, and without showing any additional polymor-
phisms were selected and included in the apple 20K
FruitBreedomics SNP array.
To verify the reliability of each candidate SNP marker, 25
apple cultivars from the germplasm collection at Fondazione
Edmund Mach, for which data of the PCR-based markers
were already available for the three genes (SCAR: Md-ACS1
and Md-ACO1, SSR: Md-PG1SSR10kd), were tested with the
apple 20K FruitBreedomics SNP array. SNP alleles were then
Tree Genetics & Genomes (2016) 12:35 Page 3 of 21 35
compared with the alleles of the PCR-based markers for these
25 apple cultivars.
Development of KASP™assays and SNP genotyping
of pilot study progenies
Genotyping was carried out with the competitive allele-
specific PCR genotyping system (KASP™, LGC Genomics
Ltd., Teddington, UK) assay, which enabled SNP bi-allelic
scoring at specific loci. Different criteria were used to select
the SNPs for which the assays were developed. FBsnRvi2-
4_R590 for Rvi2, FBsnRvi4-1_K146 for Rvi4, MS8_Y124
for Rvi6, 21k14t7_R153 and 9c10T7_Y224 for Rvi15, and
Pl2-2_Y211 for Pl2 were selected according to the informa-
tion reported by Jänsch et al. (2015), which described the high
specificity of the alleles linked to these R-genes. The parents
of progenies A and B were genotyped with the 20K SNP
array. Marker FBsnRvi2-1_M417 associated to Rvi2 and in-
cluded in this SNP array was selected for the genotyping of
progenies A and B seedlings as it was fully informative in
both crosses.
The two SNPs for the Dp-fl gene (Dp-fl_SNP_104 and Dp-
fl_SNP_585, Pagliarani et al. 2016) were selected because
they are highly specific for RAA resistanceand flank the gene.
Sequences containing the SNPs were provided to LGC
Genomics Ltd for KASP™assay design (Table 1).
Leaf samples for DNA extraction were collected from
about 5-month-old seedlings grown in a glasshouse (proge-
nies A and B) or from trees in the orchard (progenies C–J)
using the LGC DNA Extraction leaf collection kit (KBS-
9370-110). Each 96-well tube collection rack was labeled with
the ID corresponding to the virtual 96-well seedling plate.
Using a puncher, one leaf disc per plant was collected and
placed into the corresponding well of the 96-well tube collec-
tion rack. Between samplings, the puncher was rinsed in dis-
tilled water and dried on a paper tissue. The leaf discs were
placed in a rack and shipped to LGC Genomics Ltd. following
manufacturer’s instruction; no freeze-drying or freezing of the
samples was necessary.
Validation of KASP assays
For the validation of the KASP™assays for the Rvi2 and Rvi6
resistance genes, subsets of progenies A and B (62 and 58
seedlings, respectively) that according to the Rvi2SNP1 and
Rvi6SNP1 markers carried either no apple scab resistance
gene, only Rvi6 or Rvi2, or both, were selected. Two
replicates of each genotype were inoculated following the
BSwiss^protocol of Caffier et al. (2015) in the quarantine
glasshouse at Agroscope with the avrRvi6 (Rvi6-virulent)
strain EU-D42 (Caffier et al. 2015), of which conidia on dried
leaves were kindly provided by R. Groenwold (Wageningen
University and Research) and V. Bus (Plant and Food
Research NZ). Progenies A and B, their parents and the con-
trol cultivars BPriscilla^(Rvi6) and BGalaxy^(susceptible)
were inoculated in two independent experiments. Young
leaves were marked with a small plastic peg and sprayed with
conidial suspensions (progeny A = 1.35 × 10
5
conidia/ml, ger-
mination rate 68 %; progeny B = 2.2 × 10
5
conidia/ml, germi-
nation rate 75 %). Plants were evaluated using the scale of
Chevalier et al. (1991) with the addition of two further classes:
stellate chlorosis (SC) and stellate necrosis (SN) symptoms,
28 (A) and 27 (B) DAI. For each genotype, the highest symp-
tom of susceptibility observed was considered as the represen-
tative score. Progenies C to J were scored for scab incidence
after natural infections in an unsprayed orchard in 2014 and
2015, and aphid incidence in 2015. Presence/absence of spor-
ulating scab lesions or typical aphid leaf roll damage (class 3
vs classes 1 and 2 according to Pagliarani et al. (2016)) were
assessed. Resistance scoring and predictions based on the
marker results were compared. When a discordance was ob-
served, the inoculated plants were double-checked with Rvi2
flanking markers OPL19 and CH02b10 (Jänsch et al. 2015)
and CH-Vf1 (Vinatzer et al. 2004, progenies A and B), or
RT1-for and RT2-rev primers (Vinatzer et al. 2001,
progenies C to J) for Rvi6. Finally, for the field-grown plants,
all the genotype/phenotype inconsistencies were also accu-
rately re-scored for the level of resistance in the orchard.
Results
Development and preliminary tests of KASP™assays
for resistance genes
KASP™assays were successfully developed for all the
targeted SNPs using all the sequences submitted to LGC
Genomics Ltd. (Table 2). Before genotyping of the large pilot
study populations, a preliminary test of the KASP™assays
was performed on the parents and on a subset of seedlings (22
seedlings for progenies A and B and 7 for progenies C to J,
Tab le 3). The marker alleles were segregating as expected, but
for specific KASP™markers, the segregation of null alleles in
specific crosses had to be hypothesized. As KASP™assays
do not distinguish between SNP homozygous for a nucleotide
(e.g., AA) and those heterozygous including a null allele (e.g.,
Anull), the presence of null allele can only be assessed observ-
ing the allelotypes segregating in a cross.
In the crosses A and B, in the two parents not carrying Rvi2
analyzed with the marker Rvi2SNP1, the allelotype Cnull in-
stead of the CC had to be postulated. This is because both
Rvi2-resistant parents showed the CA allelotype and a part
of the progenies amplified only the nucleotide A. Assuming
the presence of a null allele in the rvi2 parents (Cnull), the
observed segregation of the allelotypes and their frequency
could be explained. Allelotype AA is actually Anull,
35 Page 4 of 21 Tree Genetics & Genomes (2016) 12:35
Tab l e 1 Sequences containing the SNPs for which KASP™assays were developed
Locus SNP name Distance to R-gene Sequence containing the SNP (accession number) Reference
Rvi2 FBsnRvi2-1_M417 1.6 cM (upstream) TAGTAGTTGCATATATTTATGATTGGTGTTGCGGA-
CGCACAGGTAAGTGCTAGGTATGTT[A/
C]ATAAATTAAAGATATGAGATTGCTTCAGTTAT-
GTGAGATATGATGTGATGTATTGAAAGC
(KM104964)
Jänsch et al. 2015
Rvi2 FBsnRvi2-4 _R590 0.9 cM (upstream) TGACCAGCCAGAGGTACAGAAAGACGAGGAAA-
ATGGCRGAGATGAAGGTCGGCAAAGTTT[A/
G]RGTCRGTACTTGATTGGTGGGATTGAGACCA-
GAGGGAGTACARGTAGCTGGAATGGGAAT
(KM104984)
Jänsch et al. 2015
Rvi4 FBsnRvi4-1_K146 Cosegregates with
the gene
AGAAGGGGAATGTTGACTGGATTAAGTTTATGGA-
TTACATATT T TGG AT T CTATAT TGG T [G/
T]AACAGCTTGCKAGTGTTTAGATGCAGGTGTA-
ACATTGATGAAATATAAGTGTTATCTAAA
(KM105043)
Rvi6 MS8_Y124 ca. 0.3 cM GAAAGGAAACATCGTCATCTAGTTGAGACAGCTA-
GAACTTTATTGGTTTAATCTCATGTTCCTCACCA-
ATATTGGGTTGAAGCCTTCACGACTGCACTATAT-
CTCATTAATAAATTG[T/
C]CAATCTCTGGGATTTCACAATCACCTTAGCAA/
GTTATTATTTCATATAGTTCCAGATTACTCAAGGC-
TCG/AAAGTCTT (KM105074)
Jänsch et al. 2015
Rvi15 21k14t7_R153 0.2 cM
(downstream)
AGAGAAACCACTTATTTTAGGCTCCTATCAGTTG-
GTGAGGTAGTGGTGCAATGTGCTCTC[A/
G]AGAGCCCTAGCAAGTGATAGATGCCATTGCG-
GAGGTTGATGACGATGAATCAGTGGCATG
(KM105120)
Jänsch et al. 2015
Rvi15 9c10T7_Y224 0.7 cM (upstream) CATGGCTTTGAAGCAATCATACTACTAASTATTAG-
CCATTGAATCTTCTACTAACAAAGT[G/
T]ATCGACTGCCTACGGAATCTTCTACTATCAAA-
TTAGCACRGTAAAACGACATTGCACAAA
(KM105101)
Jänsch et al. 2015
Pl2 FBsnPl2-2_Y211 Within the gene TGTGGCTGGATGATCTCAGAGACCTGGCCTATGA-
TATCGAAGACATGTTGGACAAATTTGCCGTTAA-
AATGTTGAAGCGCATGATAGAGGGA[T/
C]GTGATCAAGCCAGCACAAGCAGGAAGGTAC-
GGAGATCATTTTATAAAGTTAAATTGAGTTTTG-
ATATGAACTCCGAAATGAAGAAGATTACGA
(KM104953)
Jänsch et al. 2015
DP-fl- Dpfl_SNP_104 Upstream, 330 kb
from Dpfl_SNP_
585
CCGGCACCAGAGCATTTACATTCCTGGATTTAGGT-
GGACTACTTCAAACTCTNTTCCTAG[T/
C]AAACGGTTTCTCAGCTAACCGCCCTCACTCG-
GTCGGACCGCCACCAAATCCTCCGAATCG
(rs796061839)
Bianco et al. 2014
Tree Genetics & Genomes (2016) 12:35 Page 5 of 21 35
Tab l e 1 (continued)
Locus SNP name Distance to R-gene Sequence containing the SNP (accession number) Reference
DP-fl- Dpfl_SNP_585 downstream,
330 kb from
Dpfl_SNP_104
ATGAAAAACAAAAGCAGGTAAATACATCTTCAA-
AATGAATTGGGGAAAAAAAAGAAGATA[T/
C]GTTTAGGCTGATTTCAAGCGAACCCTAACGA-
ACACATGAACATTTCACATGAAACCATTG
(rs796057313)
Bianco et al. 2014
ACS1 Md-ACS1SNPa
a
Within the gene CTTTGATCCCAACCACTTAGTGCTCACCGCCGGT-
GCAACTTCAGCGAATGAGACCTTTAT[T/
C]TTCTGCCTTGCTGACCCCGGCGAAGCCGTTC-
TTATTCCTACCCCATACTACCCAGGGTAC
(ss#1966650605)
This paper
ACS1 Md-ACS1SNPb
a
Within the gene GTATAAGCTCACTTTTCGTTAATGCAATTAAAAGC-
TAC TAC TAG AA CA AGTCTTCTAG CC [G /
A]GTTGCATGTCTAACTCAGCTTTTGATTATTTTT-
TTCTTACAGGCAATGGTAGATTTCATG
(ss#1966650606)
This paper
ACO1 Md-ACOSNP1
a
Within the gene ACAACCCAGGCAACGACNCATTCATCAGCCCAG-
CACCGGCAGTGCTTGAGAAGAAAACTG[A/
G]GGACGCCCCAACTTATCCCAAGTTTGTGTTT-
GATGACTACATGAAGCTGTATTCTGGCCT
(ss#1966650604)
This paper
PG1 Md-PG1SNP1
a
Within the gene ACCCAGAACATCACTATCTCGAGCTCGGTTATAG-
GAACAGGTGTGACAAATAGTTCATGA[C/
T]GCTTTCAAAAGTCATCATTTTGGTCTCTCGCG-
TTT G ATA A ATACAT ATG AT C GATC A AAC
(ss#1966650603)
This paper
The targeted SNP is indicated between squared brackets, additional SNPs in the same sequence are indicated by a slash (e.g.,/i.e., MS8_Y124). Favorable alleles (associated with resistance or improved fruit
quality) are indicated in bold
a
The names of these four SNPs on the 20K FruitBreedomics array are Md-ACS1SNPa = FEM_cg_6; ACS1SNPb = FEM_cg_5; Md-ACOSNP1 = FEM_cg_4; Md-PG1SNP1 = FEM_cg_19)
35 Page 6 of 21 Tree Genetics & Genomes (2016) 12:35
Tab l e 2 Primer sequences of KASP™assay
Locus KASP™assay
name = SNP
marker name
SNP Allele X primer for Allele Y primer for Common Primer rev Allele X Allele Y Used to
genotype cross
Rvi2 Rvi2SNP1 FBsnRvi2-1_M417 GCACAGGTAAGTGCTA-
GGTATGTTA
CACAGGTAAGTGCTAG-
GTATGTTC
TCACATCATATCTCACATA-
ACTGAAGCAAT
ACA,B,C
Rvi2 Rvi2SNP2 FBsnRvi2-4 _R590 GTTATATAATTATGGCT-
CCCAATCAATTGT
ATATAATTATGGCTCCC-
AATCAATTGC
GCACTGGGTGCAGTGTTA-
GGGAA
AGC
Rvi4 Rvi4SNP1 FBsnRvi4-1_K146 AATGTTACACCTGCAT-
CTAAACACTC
CAATGTTACACCTGCAT-
CTAAACACTA
GGAATGTTGACTGGATTA-
AGTTTATGGATT
GTC
Rvi6 Rvi6SNP1 MS8_Y124 GGTGATTGTGAAATCC-
CAGAGATTGA
GTGATTGTGAAATCCC-
AGAGATTGG
CCTTCACGACTGCACTAT-
ATCT C ATTAAT
TCAtoJ
Rvi15 Rvi15SNP1 21k14t7_R153 GGTAGTGGTGCAATGT-
GCTCTCA
GTAGTGGTGCAATGTG-
CTCTCG
CATCAACCTCCGCAATGG-
CATCTAT
AGH
Rvi15 Rvi15SNP2 9c10T7_Y224 GGAATCTTCTACTATCA-
AATTAGCACG
CGGAATCTTCTACTATC-
AAATTAGCACA
GTTACTCATCTTCTGATGC-
TTGTTAGTGTT
CTH
Pl2 Pl2SNP1 FBsnPl2-2_Y211 GCTTGTGCTGGCTTGA-
TCACA
GCTTGTGCTGGCTTGA-
TCACG
TTGCCGTTAAAATGTTGA-
AGCGCATGATA
TCB
DP-fl Dp-flSNP1 Dpfl_SNP_104 GCGGTTAGCTGAGAAA-
CCGTTTA
GCGGTTAGCTGAGAAA-
CCGTTTG
ATTCCTGGATTTAGGTGG-
ACTACTTCAAA
TCCtoJ
DP-fl Dp-flSNP2 Dpfl_SNP_585 GGTTCGCTTGAAATCA-
GCCTAAACA
GTTCGCTTGAAATCAG-
CCTAAACG
GTAAATACATCTTCAAAA-
TGAATTGGGGAA
TCCtoJ
ACO1 ACO1SNP1 Md-ACO1SNP1 TGGGATAAGTTGGGGC-
GTCCT
GGGATAAGTTGGGGCG-
TCCC
GCACCGGCAGTGCTTGA-
GAAGAA
A G B, C, F, G, J
ACS1 ACS1SNP1 Md-ACS1SNPa CCGGGGTCAGCAAGG-
CAGAAA
CGGGGTCAGCAAGGC-
AGAAG
GTGCAACTTCAGCGAATG-
AGACCTT
TCB
ACS1 ACS1SNP2 Md-ACS1SNPb ACTACTAGAACAAGTC-
TTCTAGCCG
ACTACTAGAACAAGTC-
TTCTAGCCA
CATGAAATCTACCATTGC-
CTGTAAGAAAAA
GAB,E,G,H,I
PG1 PG1SNP1 Md-PG1SNP1 AGACCAAAATGATGAC-
TTTTGAAAGCG
GAGACCAAAATGATGA-
CTTTTGAAAGCA
GGTTATAGGAACAGGTGT-
GACAAATAGTT
C T A, B, D, G, I
Tree Genetics & Genomes (2016) 12:35 Page 7 of 21 35
Tab l e 3 Allelotypes of the parents of the crosses and relative progenies
Cross Parents R-gene(s) Rvi2SNP1
(A
a
)
Rvi2SNP2
(A
a
)
Rvi4SNP1
(T
a
)
Rvi6SNP1
(T
a
)
Rvi15SNP1
(G
a
)
Rvi15SNP2
(T
a
)
Pl2SNP1
(T
a
)
Dp-flSNP1
(C
a
)
Dp-flSNP2
(T
a
)
ACO1SNP1 ACS1SNP1 ACS1SNP2 PG1SNP1
AACW11303Rvi6rvi6 Cnull
b
––CT –––– – AA CC AA CC
ACW 18522 Rvi2rvi2,
Rvi6rvi6
CA ––CT –––– – AA CC AA TC
B ACW 13652 Rvi6rvi6,
Pl2pl2
Cnull
b
––CT ––CT ––AG CC AG TC
ACW 11567 Rvi2rvi2 CA ––CC ––CC ––AA CT AA TT
C GoldRush Rvi6rvi6,
Dp-fldp-fl
Cnull
b
GG GG CT –––CT CT AA –––
Realka’Rvi2rvi2,
Rvi4rvi4
CA GA TG CC –––CC CC AG –––
D GoldRush Rvi6rvi6,
Dp-fldp-fl
–––CT –––CT CT –––CC
Discovery scab QTL
on LG17
–––CC –––Cnull
b
CT –––TC
EFlorinaRvi6rvi6,
Dp-fldp-fl
–––CT –––CT Tnull
b
––AG –
Fuji – –––CC –––CT CC ––AA –
FFlorinaRvi6rvi6,
Dp-fldp-fl
–––CT –––CT Tnull
b
AA –––
Grenoble – –––CC –––Cnull
b
CC AG –––
GFlorinaRvi6rvi6,
Dp-fldp-fl
–––CT –––CT Tnull
b
AA –AG CC
Royal Gala – –––CC –––TT CC AG –AA TC
H GoldRush Rvi6rvi6,
Dp-fldp-fl
–––CT Nullnull
a
CC –CT CT ––AA –
GMAL2473 –––CC GA TC –CC Tnull
b
––AG –
IFlorinaRvi6rvi6,
Dp-fldp-fl
–––CT –––CT Tnull
b
––AG CC
Perleberg – –––CC –––TT CT ––AG TC
J GoldRush Rvi6rvi6,
Dp-fldp-fl
–––CT –––CT CT AA –AA –
White Angel – –––Tnull
a
–––Cnull
b
CT AG –AG –
a
Allele associated to the resistance
b
Presence of the Bnull^allele deduced from the segregation of the allelotypes in the progeny
35 Page 8 of 21 Tree Genetics & Genomes (2016) 12:35
allelotype CC is a mix of CC and Cnull, while allelotype CA is
correct. In progeny B, the frequency of the three groups Anull,
CC/Cnull and CA were 25.6, 48.5, and 25.9 %, respectively,
and thus fitting with the 1:2:1 expected frequency distribution.
Also, in progeny J, the segregation of a null allele was
predicted. Here, part of the progeny amplified only the
nucleotide C when tested with Rvi6 marker Rvi6SNP1.
Given the amplification of CT for BGoldRush^and of
only T for BWhite Angel^, the allelotype of BWhite
Angel^must be Tnull. In fact, only this later allelotype
can explain the amplification of only the C nucleotide of
part of the progeny. The observed frequencies of the three
groups CT, TT/Tnull, and Cnull were 30, 48, and 22 %,
respectively, and therefore fitted the 1:2:1 expected fre-
quency distribution.
The seedlings of progeny H, tested with the marker
Rvi15SNP1 (Rvi15) amplified only the nucleotide G or
A. Considering the parents (no amplification for
GoldRush and GA for GMAL2473), it is conceivable
that BGoldRush^has actually two null alleles. Indeed,
the allelotypes are Anull and Gnull, and they were pres-
ent in 43 and 57 % of the seedlings, respectively.
Null alleles were also detected for Dp-fl markers: Dp-
flSNP1 has a null allele in BDiscovery^,BGrenoble^,
and BWhite Angel^and thus, the allelotype correspond-
ing to the amplification of the nucleotide T in their
crosses is actually Tnull; the Dp-flSNP2 has a null al-
lele in BFlorina^and GMAL2473 and the allelotype
corresponding to the amplification of the nucleotide C
in their populations is actually Cnull.
Validation of KASP™assays for resistance genes
In order to validate the predictions using the SNP markers for
Rvi2 and Rvi6 (Rvi2SNP1 and Rvi6SNP1), a subset of prog-
eny plants of progenies A (62 genotypes) and B (58 geno-
types) were inoculated with the avrRvi6 (Rvi6-virulent) apple
scab isolate EU-D42. The inoculations were successful as
sporulating lesions (class 4) were observed on the susceptible
control BGalaxy^,theRvi6-resistant BPriscilla^,aswellasthe
parents of these two crosses, ACW 11303 and ACW 13652,
carrying Rvi6 (data not shown). In contrast, both parents car-
rying Rvi2, ACW 18522, and ACW 11567, showed no visible
symptoms (class 0). The correlation between marker predic-
tion and assessed resistance in the progeny plants of the two
crosses was consistent in 115 cases out of 120 (Table 4). The
plantlets of the five genotypes showing discordant results
were molecularly checked with the SCAR (OPL19) and the
SSR (CH02b10 and CH-Vf1) markers linked to Rvi2 and
Rvi6. In three cases, the predictions carried out with these
latter markers were concordant with their scab scoring. After
the correction of the predicted genotype of these three seed-
lings, all the genotypes predicted to carry Rvi2 resistance
(alone or in combination with Rvi6) were shown to be scab
resistant (Table 4), while all genotypes with Rvi6 or no scab
resistance, with one exception for each group, were scored as
scab susceptible (class 4).
The prediction carried out with the Rvi6SNP1 marker was
also validated using scab data of progenies C to J. Predictions
perfectly fitted with the phenotypic scab observations for
progenies D, E, and G, as scab was never observed on
Tabl e 4 Distribution of seedlings
of progenies A and B according to
the scab symptoms 28 and 27
DAI, respectively, and their apple
scab resistance genes setup
predicted by SNP markers
Scab class/symptom 0
(%)
1
(%)
2
(%)
SC
(%)
SN
(%)
3a
(%)
3b
(%)
4
(%)
Scab gene
combination
Number of plants (percent
of total plants)
Cross A
Rvi2,Rvi6 28 86 0 4 4 7 0 0 0
Rvi2 9 44 0 11 11 33 0 0 0
Rvi6 11 0000 90091
None 14 0 0 0 700093
Cross B
Rvi2,Rvi6 26 58 0 0 0 42 0 0 0
Rvi2 12 25 0 0 0 75 0 0 0
Rvi6 4 0 0 0 0 0 0 0 100
None 16 0 0 0 0 0 0 0 100
Classes 1, 2, 3a, 3b and 4 according to Chevalier et al. (1991); classes SC and SN: stellate chlorosis and necrosis.
Only class 4 plants are considered as susceptible. Thegenotype of three seedlings (2 resp. 1 of cross A and B) have
been modified according the results obtained with the SCAR and SSR additional analysis. Italic numbers are the
two plants showing resistance but that according to the marker predictions (SCAR and SSRs included) should be
susceptible to the inoculum used
Tree Genetics & Genomes (2016) 12:35 Page 9 of 21 35
genotypes carrying Rvi6 and vice versa (Table 5). A few in-
consistencies were found in the other progenies. A total of 4
(9 %), 12 (13 %), 3 (3 %), and 26 (32 %) seedlings in crosses
F, H, I, and J, respectively, were predicted to not carry Rvi6
and were not showing scab. Finally, only one genotype of
cross J was predicted to carry Rvi6, but showed scab lesions
in the orchard. All the genotypes showing inconsistencies
were tested with SCAR marker RT1-for/RT2-rev located
within the Rvi6 gene and none of them resulted to be a recom-
binant genotype for the presence or absence of this major
gene.
Validation of the markers for Rvi4 (Rvi4SNP1) and Rvi15
(Rvi15SNP1 and 2) was performed in crosses C and H, re-
spectively. In four cases out of 15 (progeny C) and in 2 cases
out of 21 (progeny H), despite the predicted presence of either
Rvi4 or Rvi15, mild scab symptoms were observed. Also, the
second marker for Rvi2 (Rvi2SNP2) was validated with or-
chard data of progeny C. All four genotypes predicted to only
carry Rvi2 resulted to be resistant.
Two K ASP
TM
assays were developed for the Dp-fl resis-
tance gene (Dp-flSNP1 and Dp-flSNP2). Predictions done by
these specific SNP markers were validated with orchard data.
The RAA infestation was relatively low and only the 35 % of
the genotypes predicted to be susceptible (134 out of 390),
showed infestation in the orchard (Table 6). Nevertheless,
the predictions done with the two markers are reliable as in
only the 2 % of the genotypes predictedto carry the Dp-fl gene
(6 out of 345) RAA, infestation was observed. Moreover,
these six genotypes were re-checked in the orchard after the
genotyping and their aphid infestation was actually very low.
Their scoring resulted the borderline between class 2 (tolerant)
and 3 (susceptible).
Pl2SNP1, ACS1SNP1 and 2, ACO1SNP1, and PG1SNP1
KASP™assays (for the identification of SNP markers asso-
ciated to Md-ACO1,Md-ACS1,andMd-PG1 used to develop
these assays see BIdentification of SNP markers associated to
Md-ACO1,Md-ACS1, and Md-PG1^Section) were finally
validated comparing the allelotypes obtained from the analy-
sis of all parents of the pilot studies progenies with these
assays and those obtained by the 20K FruitBreedomics SNP
Illumina array (Bianco et al. 2014). In all cases, no mismatch
was observed.
Identification of SNP markers associated to Md-ACO1,
Md-ACS1,andMd-PG1
For the three genes, full-length sequence alignment identified
a total of 21 SNPs: 3 for Md-ACS1, 4 for Md-ACO1,and14
for Md-PG1. To verify the consistency between these new
markers with regard to the ones already known (Md-ACS1,
Md-ACO1, and Md-PG1SSR10kd), each SNP was compared
with the PCR-based marker allelotype within the group of 25
apple cultivars assessed. For Md-ACS1,SNPsMd-
ACS1SNPa and Md-ACS1SNPb were identified as being
highly consistent with the SCAR marker allelotype and char-
acterized by [T/C] and [A/G] polymorphisms, respectively
(Table 7). However, two inconsistencies were observed for
Md-ACS1SNPb in BDalinette^and BDelblush.^For Md-
ACS1SNPa, SNP allele T was associated with the allele
known as Md-ACS1-1(513 bp), while allele C was associated
with Md-ACS1-2(651 bp). For Md-ACS1SNPb, SNP allele G
was associated with the allele known as Md-ACS1-1(513 bp),
while allele A was associated with Md-ACS1-2(651 bp), with
the above-mentioned exception of BDalinette^and
BDelblush^(Table 7). The SNP selected for Md-ACO1,name-
ly Md-ACO1SNP1, instead, was represented by a [A/G] poly-
morphism (Table 7). For Md-ACO1SNP1, the SNP alleles A
and G were linked to the Md-ACO1-2(584 bp) and Md-
ACO1-1allele (522 bp), respectively (Table 7). The SNP
Md-PG1SNP1 selected for Md-PG1 was represented by a
[T/C] polymorphism (Table 7). Since the PCR marker avail-
able for Md-PG1 is based on a SSR motif, with three alleles
(289, 292, and 298 bp) known to date (Longhi et al. 2013a),
the relationship between the microsatellite allelic configura-
tion and the SNP allelotype follows a different calling with
regard to Md-ACS1 and Md-ACO1, which are instead distin-
guished by only two alleles. SNP allele T of Md-PG1SNP1
was found to be specifically associated with the 298 bp allele
(with the exception of BTopa z^), while allele C can be asso-
ciated with both the 289 and 292 alleles. Considering this, the
Md-PG1SNP1 allelotype T/T identifies Md-PG1 298/298 ge-
notypes, allelotype T/C can be associated with both Md-PG1
289/298 and 292/298, while allelotype C/C can be defined by
three Md-PG1 allelic combinations: 289/289, 289/292, and
292/292.
Pilot studies
Progeny A was generated pollinating 2246 flowers of ACW
11303 with pollen of ACW 18522, 352 fruits were harvested,
and 2310 seeds extracted, while to generate progeny B, 1370
flowers of ACW 13652 were pollinated with pollen of ACW
11567, 323 fruits were harvested, and 2793 seeds were ex-
tracted (Table 8). For the pilot studies (Table 8), 2087 plants
for progeny A and 2578 plants for progeny B were finally
scored for scab susceptibility: 227 (10.9 %) and 511
(19.8 %) plants from progenies A and B, respectively, were
classified as susceptible (class 4), which, with the exception of
those kept as controls (54 and 36 for progenies A and B,
respectively), were removed together with the dwarf ones.
The genotyping was finally performed on 1815 and 2022
seedlings, respectively, which survived until sampling. For
1618 out of 1815 (89.1 %) and 1753 out of 2022 (86.7 %)
seedlings of progeny A and B, respectively, a complete dataset
for the resistance markers was obtained. According to the
marker predictions, 92 and 259 susceptible seedlings of
35 Page 10 of 21 Tree Genetics & Genomes (2016) 12:35
progenies A and B, respectively, were present in the two prog-
enies. Therefore, in cross A and B, 38 and 223 additional
susceptible seedlings were identified using the markers. In
total, 678 and 448 seedlings of progenies A and B,
Tabl e 5 Distribution of the
seedlings of progenies C–J
according to the presence/absence
of scab sporulation in the orchard
(data collected in June 2014 and
June 2015) and their scab resis-
tance genes setup determined by
SNP markers
Scab symptom No No Yes Yes
Scab gene combination Number of plants (percent of total plants) N%N%
Cross C
Total 117 % 98 84 19 16
None 15 13 0 0 15 13
Rvi2 4
a
34300
Rvi4 15 13 11 9 4 3
Rvi6 22 19 22 19 0 0
Rvi4+Rvi6 14 12 14 12 0 0
Rvi2+Rvi4 13 11 13 11 0 0
Rvi2+Rvi6 989800
Rvi2+Rvi4+Rvi6 25 21 25 21 0 0
Cross D
Total 174 % 169 97 5 3
None 5 3 0 0 5 3
Rvi6 169 97 169 97 0 0
Cross E
Total 43 % 40 95 3 5
None 3 7 0 0 3 7
Rvi6 40 93 40 93 0 0
Cross F
Total 44 % 36 82 8 18
None 12 27 4 9 8 18
Rvi6 32 73 32 73 0 0
Cross G
Total 94 % 92 98 2 2
None 2 2 0 0 2 2
Rvi6 92 98 92 98 0 0
Cross H
Total 90 % 82 91 8 9
None 18 20 12 13 6 7
Rvi6 32 36 32 36 0 0
Rvi15 21 23 19 21 2 2
Rvi6+Rvi15 19
b
21 19 21 0 0
Cross I
Total 92 % 87 95 5 5
None 8 9 3 3 5 5
Rvi6 84 91 84 91 0 0
Cross J
Total 81 % 64 79 17 21
None 42 52 26 32 16 20
Rvi6 39 48 38 47 1 1
a
One plant is a recombinant between the two SNPs flanking the Rvi2 gene. It is possible to deduce that the
recombination occurred after the gene because the plant is resistant
b
One plant is a recombinant between the two SNPs flanking the Rvi15 gene. It is possible to deduce that the
recombination occurred after the gene because the plant is resistant
Tree Genetics & Genomes (2016) 12:35 Page 11 of 21 35
respectively, carried the Rvi2 and Rvi6 pyramid. A total of 225
seedlings of progeny A carried Rvi2 and Rvi6 pyramided as
well as Rvi6 in the homozygous state. In progeny B, 223
seedlings carried the Rvi2,Rvi6,andPl2 pyramid.
All seedlings with pyramided R-genes (629 in prog-
eny A and 193 in progeny B) and with a complete
dataset for the fruit quality trait markers and that sur-
vived until the genotyping data were available, have
beenpottedandgrowninanopenairfieldtobefinally
evaluated in autumn for growth characteristics and pow-
dery mildew incidence. One hundred and sixty-eight and
68 seedlings with pyramided resistance genes of proge-
nies A and B, respectively, were selected and planted in
an orchard for further evaluation.
A total of 1023 plants from progenies C–J were scored for
scab and aphid resistance. In total, 795 plants were genotyped
and for 734 of them, a complete dataset for the resistance
markers was obtained (Table 9).
Tabl e 6 Distribution of the
progenies C–J according to the
presence/absence of aphid leaf
roll damages in the field (data
collected in June 2014 and June
2015) and their RAA resistance
gene setup determined by SNP
markers
Aphid symptom No No Yes Yes
Aphid gene combination Number of plants (percent of total plants) N%N%
Cross C
Total 117 % 114 97 3 3
None 57 49 54 46 3 3
Dp-fl 60
a
51 60 51 0 0
Cross D
Total 174 % 170 98 4 2
None 93 53 90 52 3 2
Dp-fl 81
b
47 80 46 1 1
Cross E
Total 43 % 25 58 18 42
None 27 63 9 21 18 42
Dp-fl 16 37 16 37 0 0
Cross F
Total 44 % 30 68 14 32
None 22 50 8 18 14 32
Dp-fl 22
b
50 22 50 0 0
Cross G
Total 94 % 60 64 34 36
None 47 50 14 15 33 35
Dp-fl 47
c
50 46 49 0 0
Cross H
Total 90 % 55 61 35 39
None 51 57 20 39 31 34
Dp-fl 39 43 36 92 3 3
Cross I
Total 92 % 60 65 32 35
None 47 51 19 21 28 30
Dp-fl 45
d
49 41 45 2 2
Cross J
Total 81 % 77 95 4 5
None 46 57 42 52 4 5
Dp-fl 35 43 35 43 0 0
a
One plant is a recombinant between the two SNPs flanking the Dp-fl gene
b
Two plants are recombinants between the two SNPs flanking the Dp-fl gene
c
One plant is a recombinant between the two SNPs flanking the Dp-fl gene. It is possible to deduce that the
recombination occurred before the gene because the plant is susceptible to RAA
d
Two plants are recombinants between the two SNPs flanking the Dp-fl gene. It is possible to deduce that the
recombination occurred before the gene because the plants are susceptible to RAA
35 Page 12 of 21 Tree Genetics & Genomes (2016) 12:35
The seedlings with pyramided R-genes for scab resistance
were 25 (21 %) in progeny C (Rvi2,Rvi4,andRvi6)and19
(21 %) in progeny H (Rvi6 and Rvi15)(Table9). No scab was
observed at any time on genotypes with pyramided R-genes in
the orchard. Considering scab and RAA resistance together,
14 (12 %) genotypes in progeny C had the full set of
pyramided R-genes (Rvi2,Rvi4,Rvi6,andDp-fl)and11
(12 %) genotypes in progeny H (Rvi6,Rvi15,andDp-fl).
For the other progenies, from D to J, the plants with the Dp-
fl and Rvi6 pyramid were 78 (45 %), 16 (37 %), 16 (36 %), 46
(49 %), 40 (44 %), and 18 (22 %), respectively (Table 9).
For the three genes related to fruit quality traits (Md-ACS1,
Md-ACO1,andMd-PG1), the first step was to define the
allelotypes of the parental cultivars based on the apple 20K
FruitBreedomics SNP array results. In progeny A, both par-
ents were homozygous AA for the Md-ACS1SNPb and Md-
ACO1SNP1 and CC for Md-ACS1SNPa SNP markers.
However, marker Md-PG1SNP1 was heterozygous in
ACW18522 (CT) and homozygous (CC) in ACW11303.
Therefore, the genotyping was performed only with the
KASP™assay PG1SNP1. In contrast, the markers for all
three fruit quality traits were polymorph in progeny B
(Table 3).
Concerning progenies C to J, Md-ACS1SNPb was hetero-
zygous (AG) in BFlorina^,BPerleberg^,andBWhite Angel^;
Md-ACO1SNP1 was heterozygous (AG) in BRealka^,
BGrenoble^,BRoyal Gala^,andBWhite Angel^;andMd-
PG1SNP1 was homozygous (CC) in BGoldRush^and
BFlorina^and heterozygous (TC) in BDiscovery^,BRoyal
Gala^,andBPerleberg^. This meant that for Md-PG1 no seed-
ling can be homozygous for the unfavorable allele T. For these
three genes, the genotyping was extended to the whole fami-
lies between heterozygous parents (Table 4). As all the parents
of the pilot studies, BWhite Angel^was tested with the func-
tional SCAR marker Md-ACS1 and the genotyping results
were compared with those obtained for Md-ACS1SNPb.
BWhite Angel^was homozygous for the 513 bp allele for
Md-ACS1 but heterozygous [A/G] for Md-ACS1SNPb (data
not shown). Due to this discrepancy, this marker could not be
used for selection in progeny J.
Tabl e 7 Comparison of SCAR or SSR with SNP markers for the Md-ACS1,Md-ACO1,andMd-PG1 genes of 25 apple genotypes
CVS Md-ACS1 Md-ACS1 Md-ACS1 CVS Md-ACO1 Md-ACO1 CVS Md-PG1
SSR
10kd Md-PG1
SNPa SNPb SNP1 SNP1
Calamari 513 513 TT GG Ariane 584 584 AA Ambrosia 298 298 TT
Limoncini 513 513 TT GG Ariwa 584 584 AA Gloster 298 298 TT
Ariwa 513 651 TC AG Calamari 584 584 AA Braeburn 289 298 TC
Braeburn 513 651 TC AG Cripps Pink 584 584 AA Calamari 289 298 TC
Cripps Pink 513 651 TC AG Dalinette 584 584 AA Dalinette 292 298 TC
Dalinette 513 651 TC AA Dalitron 584 584 AA Discovery 289 298 TC
Delblush 513 651 TC AA Delblush 584 584 AA Gold Chief 292 298 TC
Florina 513 651 TC AG Discovery 584 584 AA Golden Delicious 292 298 TC
Golden Delicious 513 651 TC AG Florina 584 584 AA Golden Orange 289 298 TC
Golden Orange 513 651 TC AG Gold Delicious 584 584 AA Limoncini 289 298 TC
Granny Smith 513 651 TC AG Golden Orange 584 584 AA Nicoter 289 298 TC
Ligol 513 651 TC AG Granny Smith 584 584 AA Topaz 292 298 CC
Scifresh 513 651 TC AG Limoncini 584 584 AA Ananas Reinette 289 292 CC
Topaz 651 651 CC AA Ligol 584 584 AA Ariane 289 292 CC
Ambrosia 651 651 CC AA Pilot 584 584 AA Ariwa 289 289 CC
Ananas Reinette 651 651 CC AA Pinova 584 584 AA Cripps Pink 292 292 CC
Ariane 651 651 CC AA Topaz 584 584 AA Dalitron 289 292 CC
Dalitron 651 651 CC AA Ambrosia 522 584 AG Delblush 292 292 CC
Discovery 651 651 CC AA Ananas Reinette 522 584 AG Florina 289 289 CC
Fuji 651 651 CC AA Braeburn 522 584 AG Fuji 289 289 CC
Gloster 651 651 CC AA Gloster 522 584 AG Granny Smith 292 292 CC
Gold Chief 651 651 CC AA Gold Chief 522 584 AG Ligol 292 292 CC
Nicoter 651 651 CC AA Nicoter 522 584 AG Pilot 289 289 CC
Pilot 651 651 CC AA Scifresh 522 584 AG Pinova 289 292 CC
Pinova 651 651 CC AA Fuji 522 522 GG Scifresh 289 292 CC
Underlined are the SNP alleotypes not matching with the SCAR/SSR marker for the same locus
Tree Genetics & Genomes (2016) 12:35 Page 13 of 21 35
Out of the 168 resistant genotypes of progeny A
planted in the orchard, 86 genotypes carried the favor-
able allele of marker Md-PG1SNP1. In cross B, only 3
out of the 68 resistant genotypes planted in the orchard
carried the best allelic combination of the markers for
the genes Md-ACS1,Md-ACO1,andMd-PG1
(Table 8).
For the progenies in the orchard, the resistant genotypes
with pyramided R-genes for scab carrying also the favorable
allele for ACS1SNP2, ACO1SNP1, and PG1SNP1 were 14
(10 %) in progeny C (Rvi2, Rvi4, Rvi6, and Md-ACO1)and11
(10 %) in H (Rvi6, Rvi15, and Md-ACS1)(Table9). When the
Dp-fl gene was included in the analysis, genotypes with
pyramided R-genes for scab and RAA, and with favorable
Tabl e 8 Statistics of the pilot studies progenies A and B
Progeny AB
No. of pollinated flowers 2246 1370
No. of harvested fruits 352 323
No. of seeds gained 2310 2793
No. seedlings inoculated with scab 2087 2578
No. (%) of phenotypically susceptible seedlings (class 4) 227 (10.9) 511 (19.8)
No. of susceptible seedling kept as control 54 36
No. of dwarf seedlings removed and seedlings that died
before sampling for genotyping
99 81
No. of genotyped seedlings 1815 2022
No. (%) of seedlings with complete SNP data for all
resistance genes segregating in the cross (2 resp 3
in cross A and B)
1618 (89.1) 1753 (86.7)
No. of susceptible seedlings according to SNP data 92 259
No. of genotypes carrying Rvi2rvi2 and Rvi6Rvi6 or Rvi6rvi6 678 448
No. of genotypes carrying Rvi2rvi2 and Rvi6/Rvi6 225 n.a.
No. of genotypes carrying Rvi2rvi2,Rvi6rvi6,andPl2pl2 n.a. 223
No. of genotypes with full dataset for R genes and fruit
quality markers
1542 1603
No. (%) of genotypes carrying allelotype 1 of Md-PG1 (TC) 833 (54) 809 (50.5)
No. (%) of genotypes carrying allelotype 2 of Md-PG1 (CC) 709 (46) 794 (49.5)
No. (%) of genotypes carrying allelotype 1
of Md-ACO1 (GA)
n.a. 824 (51.4)
No. (%) of genotypes carrying allelotype 2
of Md-ACO1 (AA)
n.a. 779 (48.6)
No. (%) of genotypes carrying allelotype 1 of Md-ACS1 (GA) n.a. 901 (56.2)
No. (%) of genotypes carrying allelotype 2 of Md-ACS1 (AA) n.a. 702 (43.8)
No. of Rvi2 and Rvi6 resistant genotypes with fulldata set for
PG1SNP1 potted and phenotypically evaluated in the
container field for growth characteristics
629
No. of Rvi2 and Rvi6 resistant genotypes with good growth
characteristic phenotypically selected
168
No. of Rvi2 and Rvi6 resistant genotypes best combination of
alleles for fruit quality trait markers (PG1SNP1= CC)
86
No. of Rvi2 and Rvi6 resistant genotypes worst combination
of alleles for fruit quality trait markers (PG1SNP1 =CT)
82
No. of Rvi2,Rvi6,andPl2 resistant genotypes with full data
set for fruit SNPs, potted and phenotypically evaluated in
the container field for growth characteristics
193
No. of Rvi2,Rvi6,andPl2 resistant genotypes with good
growth characteristics phenotypically selected
68
No. of Rvi2,Rvi6,andPl2 resistant genotypes with best
combination of alleles for fruit quality trait markers
(ACO1SNP1= GA; ACS1SNP2 = AA; PG1SNP1 = TC)
3
No. of Rvi2,Rvi6,andPl2 resistant genotypes with worst
combination of alleles for fruit quality traits markers
(ACO1SNP1= AA; ACS1SNP2 = GA; PG1SNP1 = TT)
24
35 Page 14 of 21 Tree Genetics & Genomes (2016) 12:35
Tabl e 9 Statistics of the pilot studies crosses C to J
Cross C D E F G H I J
No. of seedlings in field 136 190 86 56 160 95 207 93
No. of seedlings genotyped 136 190 46 47 94 95 94 93
No. of phenotypically susceptible
seedlings kept in the study (class 4)
21 6 2 10 2 8 5 18
Percent of phenotypically susceptible
seedlings kept in the study (class 4)
15 3 4 21 2 8 5 19
No. (%) of seedlings with complete
SNP data for all resistance genes
117 (86) 174 (92) 43 (93) 44 (94) 94 (100) 90 (95) 91 (97) 81 (87)
No. (%) of seedlings with complete
dataset that according SNP are
predicted to be apple scab
susceptible
15 (13) 5 (3) 3 (7) 12 (27) 2 (2) 18 (20) 8 (9) 42 (52)
No. (%) of seedlings with complete
dataset that according to scab
infections are susceptible
19 (16) 5 (3) 3 (7) 8 (18) 2 (2) 8 (9) 5 (5) 17 (21)
No. (%) of genotypes carrying Rvi2
and Rvi6
9(8)
No. (%) of genotypes carrying Rvi2,
Rvi4,andRvi6
25 (21)
N° (%) of genotypes carrying Rvi6
and Rvi15
19 (21)
No. (%) of genotypes carrying Rvi6
and Dp-fl
32 (27) 78 (45) 16 (37) 16 (36) 46 (49) 24 (27) 40 (44) 18 (22)
No. (%) of genotypes carrying Rvi2,
Rvi6,andDp-fl
18 (15)
No. (%) of genotypes carrying Rvi2,
Rvi4,Rvi6,andDp-fl
14 (12)
No. (%) of genotypes carrying Rvi15,
Rvi6,andDp-fl
11 (12)
No. of genotypes with full dataset for
R genes and quality markers
117 174 42 43 92 90 90 81
No. (%) of genotypes carrying
allelotype 1 of Md-PG1 (CC)
81 (47) 47 (51) 39 (43)
No. (%) of genotypes carrying
allelotype 2 of Md-PG1 (TC)
93 (53) 45 (49) 51 (57)
No. (%) of genotypes carrying
allelotype 1 of Md-ACO1 (GA)
46 (39) 18 (42) 52 (57) 34 (42)
No. (%) of genotypes carrying
allelotype 2 of Md-ACO1 (AA)
71 (61) 25 (58) 40 (43) 47 (58)
No. (%) of genotypes carrying
allelotype 1 of Md-ACS1 (AA)
24 (57) 52 (57) 51 (57) 27 (30) 49 (60)
No. (%) of genotypes carrying
allelotype 2 of Md-ACS1 (GA)
19 (45) 40 (43) 39 (43) 33 (37) 32 (40)
No. (%) of genotypes carrying
allelotype 3 of Md-ACS1 (GG)
30 (33)
No. of Rvi6 resistant genotypes best
combination of alleles for quality
traits
25 79 24 11 10 31 9 16
No. of Rvi6 resistant genotypes worst
combination of alleles for quality
traits
45 90 16 20 4 20 12 23
No. of Rvi2 and Rvi6 resistant
genotypes best combination of
alleles for quality traits
16
No. of Rvi2 and Rvi6 resistant
genotypes worst combination of
alleles for quality traits
18
No. of Rvi2,Rvi4,andRvi6 resistant
genotypes best combination of
alleles for quality traits
14
Tree Genetics & Genomes (2016) 12:35 Page 15 of 21 35
alleles of the quality trait markers were nine (8 %) in progeny
C(Rvi2, Rvi4, Rvi6, and Md-ACO) and six (7 %) in H (Rvi6,
Rvi15, and Md-ACS)(Table9). Considering only the major R-
gene for scab, 91 plants out of the 291resistant genotypes with
Rvi6 and Dp-fl pyramided, are carrying the favorable alleles of
the fruit quality markers (15, 35, 10, 3, 6, 12, 3, and 8 in
crosses C–J, respectively, Table 9).
Discussion
Development and validation of KASP™assays
KASP™assays were successfully designed based on the
sequences submitted to the genotyping company. As
confirmed from the amplification of the expected alleles
of the parents, and later confirmed during the validation
process, the assays were highly specific for all the loci
of interest. These results are in agreement with evalua-
tion of the accuracy of genotype calls made by next-
generation sequencing (NGS) techniques, such as
genotyping by sequencing (GBS), by testing a small
subset of binary SNPs (270 SNPs) using an independent
KASP genotyping platform. In sweet potato, for in-
stance, the genotype calls of GBS were found to be
over 99 % consistent with KASP when scored as ho-
mozygous or heterozygous (Uitdewilligen et al. 2013).
Since KASP™assays do not distinguish between SNP
homozygous for a nucleotide and those heterozygous includ-
ing a null allele, a proper interpretation of the genotyping
results was necessary. In fact, null alleles had to be hypothe-
sized for markers linked to Rvi2,Rvi6,Rvi15,andDp-fl in
some of the progenies. Also, a double null allele was hypoth-
esized in BGoldRush^for marker Rvi15SNP1 (Rvi15).
Tabl e 9 (continued)
Cross C D E F G H I J
No. of Rvi2,Rvi4,andRvi6 resistant
genotypes worst combination of
alleles for quality traits
11
No. of Rvi15 and Rvi6 resistant
genotypes best combination of
alleles for quality traits
11
No. of Rvi15 and Rvi6 resistant
genotypes worst combination of
alleles for quality traits
7
No. of Rvi6 and Dp-fl resistant
genotypes with best combination of
alleles for quality traits
15 35 10 3 6 12 3 8
No. of Rvi6 and Dp-fl resistant
genotypes with worst combination
of alleles for quality traits
18 43 7 13 1 12 3 10
No. of Rvi2,Rvi6,andDp-fl resistant
genotypes with best combination of
alleles for quality traits
11
No. of Rvi2,Rvi6,andDp-fl resistant
genotypes with worst combination
of alleles for quality traits
8
No. of Rvi2,Rvi4,Rvi6,andDp-fl
resistant genotypes with best
combination of alleles for quality
traits
9
No. of Rvi2,Rvi4,Rvi6,andDp-fl
resistant genotypes with worst
combination of alleles for quality
traits
6
No. of Rvi15,Rvi6,andDp-fl resistant
genotypes with best combination of
alleles for quality traits
6
No. of Rvi15,Rvi6,andDp-fl resistant
genotypes with worst combination
of alleles for quality traits
5
Markers used to genotype each cross are reported in Table 2. Marker Md-ACS1SNPb was extended in family J as in all the other families between
heterozygous parents but it was not used for MAB because it is not informative, as reported in Results and Discussion
35 Page 16 of 21 Tree Genetics & Genomes (2016) 12:35
With the wide genetic background analyzed in the present
pilot study, the presence of null alleles for specific markers for
some genotypes is not surprising. In contrast with other PCR
techniques such as SSRs (Dąbrowski et al. 2014), it is not easy
to identify the presence of null alleles by using SNP markers if
segregating progenies are not available (Cuenca et al. 2013).
Markers Rvi6SNP1 and Rvi2SNP1 have been validated by
inoculating a subset of progeny plants of crosses A and B with
the avrRvi6 apple scab isolate EU-D42. In 95.8 % of the cases,
the prediction of the markers and the observed resistance/
susceptibility of the genotypes fitted perfectly, allowing the
validation of the two markers. However, the phenotype of 5
progeny plants (out of 120) did not fit with the marker predic-
tions. These five genotypes have been further re-tested with
SCAR (OPL19) and SSR (CH02b10 and CH-Vf1) markers
linked to Rvi2 and Rvi6. The prediction of these markers fitted
with the phenotype in three out of the five cases. Considering
that SSR CH02b10 and the amplicon FBsnRvi2-1 containing
Rvi2SNP1 have been mapped by Jänsch et al. (2015)atthe
same distance from Rvi2, and that CH-Vf1 (Vinatzer et al.
2004) and MS8 (Patocchi et al. 1999) containing Rvi6SNP1
are at similar distances from Rvi6 (ca 0.1 vs 0.3 cM), the
different predictions observed for the two different types of
markers can probably not be ascribed to their genetic distance
to the R-genes, but might be due to a mixing up of budwood
while preparing the plant material for the inoculation
(grafting). In fact, while the SNP analysis was carried out
using seedling leaf material, the SCAR/SSR analysis was per-
formed with DNA from plants that were inoculated with EU-
D42. For two genotypes, the SCAR/SSR markers confirmed
the genotypes predicted by the SNP markers. Both genotypes
were expected to be susceptible but they were free of scab,
hypothesizing that they may carry unknown apple scab resis-
tances (e.g., QTLs).
Marker predictions for Rvi2 (marker Rvi2SNP2), Rvi4,
Rvi6,Rvi15,andDp-fl have been verified with orchard data
collected from crosses C to J. Using orchard data for valida-
tion is more difficult as infections or infestations may not be
homogeneous in the orchard, so data collected over several
years are necessary to draw final conclusions. Moreover, these
observations can be made only on symptomatic genotypes,
and in fact, only then we can be sure that an infection/
infestation took place. Predictions with marker Rvi6SNP1,
already validated with artificial inoculations, were also con-
firmed by orchard data of progenies C to J. Only one genotype
of cross J expected to be resistant based on the Rvi6 marker,
showed scab symptoms. This genotype/phenotype inconsis-
tency could be due to a sampling mistake. In addition, it
should be noted that no scab symptoms were observed in 19
genotypes out of 63 predicted to not carry an apple scab resis-
tance gene in progenies C to I. These genotypes may have
escaped scab infection in 2014 and 2015, or they may carry
unknown resistances (e.g., QTLs). In the White Angel-
derived progeny J, a large proportion of the seedlings (32 %)
did not show scab symptoms in spite of the absence of the
marker for Rvi6 (Table 5). It is unlikely that such a large
proportion of seedlings escaped scab infection and the ob-
served discrepancy can be explained only by the segregation
of a further unknown resistance gene. The scab resistance of
the crab apple BWhite Angel^has already been reported
(Beckerman et al. 2009) and its resistance is not related to
the Rvi6 gene as proved by molecular analysis with the RT1-
for/RT2-rev amplified marker (data not shown).
Four and two genotypes of progenies C and H, respective-
ly, predicted to carry the apple scab resistance genes Rvi4 and
Rvi15, showed scab symptoms in the orchard. The co-
segregation of Rvi4 and its marker and the use of two markers
flanking Rvi15 makes the recombination between markers and
the resistance genes highly unlikely; therefore, either a mix-up
of plant material during leaf collection or a miss-scoring of
resistance remains as possible explanation for the differences
between allelotype and phenotype. Presence of race (4) and
race (15) scab strains in the experimental orchard are unlikely
as to date such strains still have to be confirmed for Rvi15 and
are rare for Rvi4 (see www.vinquest.ch; Patocchi
unpublished).
The markers for apple aphid resistance Dp-fl were validat-
ed with orchard data. Aphid infestations in the field are de-
pending on many environmental factors, such as temperature
and rainfall. The number of plants with aphids damage gen-
erally was low, since about only one third (35 %) of the ge-
notypes predicted by the markers to be susceptible were also
infested with aphids. On the other hand, only 2 % of the
genotypes predicted by markers to be aphid resistant, were
infested with aphids. This low rate of unexpected data led us
to conclude that these markers can also be considered as
validated.
The predictions suggested by the Pl2SNP1 marker could
not be validated by phenotypic data. For this, orchard mil-
dew incidence data of a Pl2SNP1 genotyped progeny segre-
gating for Pl2 over several years are needed. This is partic-
ularly true for mildew resistance for which differences in
resistance of seedlings compared to older trees have been
reported (Dunemann et al. 2007). However, the validation
of Pl2 markers with orchard data can be hampered by the
presence of powdery mildew strains able to overcome the
Pl2 resistance (Caffier and Laurens 2005). The Pl2SNP1
marker has been developed using an SNP within the open
reading frame of the Pl2 resistance gene, for which the
specificity of the allele in coupling with the resistance has
been demonstrated (Jänsch et al. 2015). These facts, together
with the consistency of the segregation ratio of the marker
fitting the expected 1:1 ratio (225 out of 452 plants carrying
Rvi2 and Rvi6 are predicted by the marker to also carry Pl2,
Tab le 8) suggest the direct use of this marker in breeding
without further validation.
Tree Genetics & Genomes (2016) 12:35 Page 17 of 21 35
Identification of SNP markers associated with Md-ACS1,
Md-ACO1,andMd-PG1
In this work, four SNPs associated with Md-ACS1,Md-ACO1,
and Md-PG1 have been identified and exploited in pilot stud-
ies. Their allelic consistency with regard to their known PCR-
based markers was verified. Md-ACS1, Md-ACO1, and Md-
PGSSR10kd, are in fact known markers implemented in sev-
eral breeding programs (Zhu and Barritt 2008; Chagné et al.
2014, RosBREED project: www.rosbreed.org), since their
association with important fruit quality traits has been
studied and validated (Harada et al. 2000; Costa et al. 2005,
2010;Longhietal.2013a,b). For the first two genes, two
SCAR markers have been developed, Md-ACS1 and Md-
ACO1, both distinguished by two alleles. Previous studies
associated the homozygosity for allele 2 (651 bp) of Md-
ACS1 and allele 1 (522 bp) of Md-ACO1 to a low production
of ethylene, thus improving the postharvest storability of the
fruit. In the allelic comparison carried out on our validation
scheme (represented by 25 apple cultivars) allelotype Md-
ACS1-2/2 was also characterized by allelotype Md-
ACS1SNPa-CC, while allelotype Md-ACO1-1/1 was associ-
ated with Md-ACO1SNP1-GG. The perfect agreement ob-
served within the group of cultivars considered in this study
hypothesizes these two SNPs are in tight linkage disequilibri-
um (LD) with the INDEL polymorphism of the respective
PCR-based markers. Regarding Md-ACS1,SNPMd-
ACS1SNPb disagreed with the phenotypes of the two culti-
vars BDalinette^and BDelblush^, which suggests a major ge-
netic distance between this SNP polymorphism and the
INDEL. In these cultivars, the allelotypes were Md-ACS1-1/
2 and Md-ACS1SNPb-AA. Inconsistencies between this SNP
polymorphism and the INDEL were also found in BWhite
Angel,^one of the parents of cross J. In this genotype, the
two allelotypes were Md-ACS1SNPb-AG and Md-ACS1-1/1
(513/513). This observation was validated by the absence of
segregation of this latter allele in the related progeny (data not
shown). As consequence, the marker Md-ACS1SNPb could
not be used for MAB selection in this progeny.
When using this SNP marker, it is recommended to first verify
that the functional allelism agrees with the SNP variants, as was
performed for all the progenies. Once this correspondence is
verified, the markers can be considered as equivalent. This was
also confirmed by the comparison of the genotyping results of
the markers ACS1SNP1 and 2 based on the SNPs Md-
ACS1SNPa and Md-ACS1SNPb in progeny B, where a single
mismatch in 1668 cases was observed (data not shown).
Since Md-PG1 was initially defined by an SSR type of
marker, more alleles (three) were identified with respect to
the previous functional markers. For this marker, the funda-
mental criterion was assigned to the SNP allele T that corre-
sponds with allele 298 for low texture performance. As dem-
onstrated by Longhi et al. (2013a), the double dose of 298
predicts a severe texture loss during storage. In the validation
panel (Table 7), the allele T was, with the only exception of
BTop az^, always associated with the 298 allele, while the oth-
er SNP allele (C) was indistinctively associated with both the
289 and 292 alleles. It is also worth noting that Longhi et al.
(2013a) did not report any difference between these two al-
leles in terms of allelic effect, thus confirming that the C allele
can be considered as associated with the favorable alleles.
Within the entire validation scheme, one inconsistency was
found for Md-PG1SNP in BTo pa z^(Table 7). Although the
SNP allelotype is CC, this cultivar presents the alleles 292 and
298, with the latter allele being present in all other tested
cultivars and associated with the SNP allele T. This can be
due to the fact that this SNP resulted in a quite high LD with
the microsatellite (r2 = 0.35, con sidering a threshold of
r2 = 0.106, as shown in Longhi et al. 2013a), enabling, how-
ever, possible event of recombination to occur.
Taking the several allelotype scenarios for these three func-
tional markers into consideration, the most favorable allelic
condition for a valuable fruit quality (improved firmness and
storability) is represented by Md-ACS1SNPa-CC or Md-
ACS1SNPb-AA and Md-ACO1SNP1-GG and Md-
PG1SNP1-CC, respectively. In light of this, the additive effect
of these alleles is worth noting, suggesting that also heterozy-
gous individuals might show interesting phenotypes, howev-
er, with a lower predictability. What certainly needs to be
avoided is to proceed in the selection process with homozy-
gous genotypes for the unfavorable alleles: Md-ACS1SNPa-
TT or Md-ACS1SNPb-GG, Md-ACO1SNP1-AA, and Md-
PG1SNP1-TT, respectively.
Pilot studies
The pilot studies have been designed to develop and test a
MAB protocol for the screening of large apple progenies with
a relatively small number of markers and suitable also for
breeders without access to their own BDNA-laboratory.^To
grow apple seedlings in a 6 × 4 well format (four trays) in
order to match the 96 wells format of the microplates gener-
ally used by the genotyping companies proved to be a helpful
strategy for tracking samples throughout the entire genotyping
process. The seedlings had sufficient soil to grow for a period
of up to 5 months. As plant materials can be sampled as soon
as the seedlings have developed the first leaf (about 1 month
after sowing) and the time necessary for the genotyping gen-
erally does not require more than 2 months, these trays enable
growing of thousands of seedlings in a limited space, avoiding
the time-consuming labeling of single seedlings. The link be-
tween the plant samples sent for genotyping and the plant of
origin was ensured by coding the samples according to the
number of the 96-well tube collection rack and the coordinates
of the 96-well virtual plate. As DNA extraction is still the most
expensive step in MAB, and most genotyping companies use
35 Page 18 of 21 Tree Genetics & Genomes (2016) 12:35
the 96-well rack, the number of missing plants in such a rack
should be minimized. This can either be done by transplanting
seedlings in the holes of missing plants prior to sampling.
Some of the holes should also be filled in the collection rack
with leaf material of the parents of the cross. Transplanting of
the seedlings was time consuming and led to stress for the
seedlings. A solution to minimize these problems could be
to combine the trays with jiffy strips (Jiffy International AS,
Kristiansand, Norway). This solution is currently used at
Better3Fruit (Belgium, I. De Wit pers. comm.).
In progenies A and B, the objective was the identification
of seedlings combining pyramided resistances against apple
scab and powdery mildew with the good growth and fruit
quality (expressed as improved firmness and storability). As
12.5 and 25 % of apple scab susceptible seedlings were ex-
pected for cross A and B, respectively, both progenies have
been inoculated with scab and the susceptible seedlings have
been removed, which saved the costs of DNA extraction of
648 samples. According to the marker analysis in cross A and
B, 38 resp. 223 seedlings genotyped were not identified as
susceptible following inoculation (Table 8). A possible expla-
nation for the higher number of susceptible seedlings that
were not identified in cross B phenotypically could be related
to the different interpretation of the scab symptoms by the
different persons that did the first culling of the seedlings of
cross A and B as both crosses were inoculated simultaneously,
in the same place and using the same inoculum.
The pilot studies with progenies C to J were performed in
order to test a MAB protocol at a more advanced stage of
selection compared with progenies A and B. Seedlings of
these progenies, for which a pre-selection for scab resistance
had already been performed, were genotyped with the fruit
quality markers informative in the specific crosses. In this
context, the pyramiding of different R-genes for scab and
RAA resistance and favorable alleles for quality traits was
achieved. Particularly interesting was the situation in crosses
C and H where, thanks to the new markers, it was possible to
select seedlings with several R genes for scab resistance (Rvi2,
Rvi4,andRvi6 in C and Rvi15 and Rvi6 in H) plus an R-gene
for RAA resistance and the favorable alleles for fruit quality
traits. Nine and six promising seedlings in cross C and H,
respectively, were selected.
KASP™-SNP Genotyping technology allowed a robust
development of trait-specific SNP marker assays (see the suc-
cessful development of an assay for each sequence carrying a
SNP sent to the company). It is extremely flexible as any assay
is run independently from the others, and it is a relatively low-
cost system for genotyping small numbers of SNPs across
large breeding populations. Considering the needs of the pilot
studies, this cost-effective and robust technology looked to be
also ideal for breeding companies interested to initiate MAB
programs, but without the availability of laboratory for DNA
genotyping.
For progenies A and B, it was decided to genotype all the
resistant seedlings (assessed phenotypically) with the fruit
quality markers informative in the specific crosses. As the data
of these markers were finally considered only for the resistant
seedlings that passed the selection for growth characteristics
and powdery mildew in the container field, the genotyping
with the fruit quality markers could have been done only on
this material (sequential genotyping). Genotyping companies
may require for sequential genotyping that the DNA is re-
extracted from the seedlings that passed the first stages of
selection or, if done by the company, may charge costs for
the re-assembling of the DNA of these seedlings in new plates.
Genotyping costs of few samples with few SNPs may be
higher per data point than large-scale genotyping. Defining
in advance the selection strategy (i.e., the time points at which
the seedlings will be selected phenotypically and/or genotyp-
ically for specific traits/loci) that will be applied to a specific
progeny and asking quotes from the genotyping companies
for the different time points at which a defined number of
markers will be applied on an expected number of seedlings,
will permit to optimize the MAB process also from an eco-
nomical point of view.
Conclusion
The present work allowed the completion of the transition from
an MAB performed using SSR and/or SCAR markers to MAB
using SNP markers. SNP-based KASP
TM
assays have been de-
veloped and validated for the currently most used in breeding
apple scab resistance genes Rvi2,Rvi4,andRvi6 and the powdery
mildew Pl2 resistance gene. The KASP
TM
assay for Rvi15,an
apple scab resistance gene basically so far not used in breeding,
will allow to broad the genetic basis of future apple scab-resistant
cultivars, while the assays for the apple aphid resistance Dp-fl,for
the first time, allowed to start applying MAB for this trait. All
these validated assays combined with those for improved fruit
storability and firmness and the proposed genotyping pipeline
that can be applied by any apple breeders will pave the way to
amoreeffectiveMABinapple.AmoreeffectiveMABwill
allow the selection of a higher number of resistant apple cultivars
with outstanding fruit quality, increasing the chance that such
cultivars will be grown in orchards, finally making apple produc-
tion more sustainable.
Acknowledgments This work has been (partly) funded under the EU
seventh Framework Programme by the FruitBreedomics project No.
265582: Integrated approach for increasing breeding efficiency in fruit
tree crop. The views expressed in this work are the sole responsibility of
the authors and do not necessary reflect the views of the European
Commission. The authors thank Simone Schütz, Rolf Blapp, Jürgen
Krauss, and René Total for assistance during the pilot studies performed
at Agroscope in Wädenswil.
Tree Genetics & Genomes (2016) 12:35 Page 19 of 21 35
References
Baumgartner IO, Patocchi A, Frey JE, Peil A, Kellerhals M (2015)
Breeding elite lines of apple carrying pyramided homozygous resis-
tance genes against apple scab and resistance against powdery mil-
dew and fire blight. Plant Mol Biol Rep 33:1573–1583
Beckerman J, Chatfield J, Draper E (2009) A 33-year evaluation of resis-
tance and pathogenicity in the apple scab–crabapples pathosystem.
HortSci 44:599–608
Bianco L, Cestaro A, Sargent DJ, Banchi E, Derdak S, Di Guardo M,
Troggio M (2014) Development and validation of a 20K single
nucleotide polymorphism (SNP) whole genome genotyping array
for apple (Malus ×domestica Borkh). PLoS ONE 9(10), e110377
Bus VG, Esmenjaud D, Buck E, Laurens F (2009) Application of genetic
markers in rosaceous crops, Genetics and Genomics of Rosaceae.
Springer, New York, pp 563–599
Caffier V, Laurens F (2005) Breakdown of Pl2, a major gene of resistance
to apple powdery mildew, in a French experimental orchard. Plant
Pathol 54:116–124
Caffier V, Patocchi A, Expert P, Bellanger MN, Durel CE, Hilber-Bodmer
M, Broggini GAL, Groenwold R, Bus V (2015) Virulence charac-
terization of Venturia inaequalis reference isolates on the differential
set of Malus hosts. Plant Dis 99:370–375
Chagné D, Crowhurst RN, Troggio M, Davey MW, Gilmore B, Lawley
C, Vanderzande S, Hellens RP, Kumar S, Cestaro A, Velasco R,
Main D, Rees JD, Iezzoni A, Mockler T, Wilhelm L, Van de Weg
E, Gardiner SE, Bassil N, Peace C (2012) Genome-wide SNP de-
tection, validation, and development of an 8K SNP array for apple.
PLoS One 7(2), e31745
Chagné D, Dayatilake D, Diack R, Oliver M, Ireland H, Watson A,
Gardiner SE, Johnston JW, Schaffer RJ, Tustin S (2014) Genetic
and environmental control of fruit maturation, dry matter and firm-
ness in apple (Malus xdomestica Borkh). Horm Res. doi:10.1038/
hortres.2014.46
Chevalier M, Lespinasse Y, Renaudin S (1991) A microscopic study of
the different classes of symptoms coded by the Vf gene in apple for
resistance to scab (Venturia ineaqualis). Plant Pathol 40:249–256
Collard BCY, Mackill DJ (2008) Marker-assisted selection: an approach
for precision plant breeding in the twenty-first century. Philos Trans
R Soc Lond B Biol Sci 363:557–572
Costa F, Stella S, Van de Weg WE, Guerra W, Cecchinel M, Dallavia J,
Koller B, Sansavini S (2005) Role of the genes Md-ACO1 and Md-
ACS1 in ethylene production and shelf life of apple (Malus
domestica Borkh.). Euphytica 141:181–190
Costa F, Peace CP, Stella S, Serra S, Musacchi S, Bazzani M, Sansavini S,
Van de Weg WE (2010) QTL dynamics for fruit firmness and soft-
ening around an ethylene-dependent polygalacturonase gene in ap-
ple (Malus×domestica Borkh.). J Expt Botany 61:3029–3039
Costa F, Cappellin L, Longhi S, Guerra W, Magnago P, Porro D,
Soukoulis C, Salvi S, Velasco R, Biasioli F, Gasperi F (2011)
Assessment of apple (Malus ×domestica Borkh.) fruit texture by a
combined acoustic-mechanical profiling strategy. Postharvest Biol
Technol 61:21–28
Costa F, Cappellin L, Fontanari M, Longhi S, Guerra W, Magnago P,
Gasperi F, Biasioli F (2012) Texture dynamics during postharvest
cold storage ripening in apple (Malus xdomestica Borkh.).
Postharvest Biol Technol 69:54–63
Cova V, Bandara NL, Liang W, Tartarini S, Patocchi A, Troggio M, Velasco
R, Komjanc M (2015) Fine mapping of the Rvi5 (Vm) apple scab
resistance locus in the ‘Murray’apple genotype. Mol Breeding 35:1–12
Cuenca J, Aleza P, Navarro L, Ollitrault P (2013) Assignment of SNP
allelic configuration in polyploids using competitive allelespecific
PCR: application to citrus triploid progeny. Ann Bot. doi:10.1093/
aob/mct032
Dąbrowski MJ, Pilot M, Kruczyk M, Żmihorski M, Umer HM, Gliwicz J
(2014) Reliability assessment of null allele detection: inconsis-
tencies between and within different methods. Mol Ecol Resour
14:361–373
Dunemann F, Peil A, Urbanietz A, Garcia‐Libreros T (2007) Mapping of
the apple powdery mildew resistance gene Pl1 and its genetic asso-
ciation with an NBS‐LRR candidate resistance gene. Plant Breed
126:476–481
Guan Y, Peace C, Rudell D, Verma S, Evans K (2015) QTLs detected for
individual sugars and soluble solids content in apple. Mol Breeding
35:135
Harada T, Sunako T, Wakasa Y, Soejima J, Satoh T, Niizeki M (2000) An
allele of the 1-aminocyclopropane-1-carboxylate synthase gene (Md-
ACS1) accounts for the low level of ethylene production in climacteric
fruits of some apple cultivars. Theor Appl Genet 101:742–746
Jänsch M, Broggini GAL, Weger J, Bus VGM, Gardiner SE, Bassett H,
Patocchi A (2015) Identification of SNPs linked to eight apple dis-
ease resistance loci. Mol Breeding 35:45. doi:10.1007/s11032-015-
0242-4
Laurens F, Durel C-E, Patocchi A, Peil A, Salvi S, Tartarini S, Velasco R,
van der Weg E (2010) Review on apple genetics and breeding
programmes and presentation of a new European initiative to in-
crease fruit breeding efficiency. J Fruit Sci 27:102–107
Longhi S, Cappellin L, Guerra W, Costa F (2013a) Validation of a func-
tional molecular marker suitable for marker-assisted breeding for
fruit texture in apple (Malus xdomestica Borkh.). Mol Breeding
32:841–852
Longhi S, Hamblin MT, Trainotti L, Peace CP, Velasco R, Costa F
(2013b) A candidate gene based approach validates Md-PG1 as
the main responsible for a QTL impacting fruit texture in apple
(Malus xdomestica Borkh.). BMC Plant Biol 13:37
Padmarasu S, Sargent DJ, Jaensch M, Kellerhals M, Tartarini S, Velasco
R, Troggio M, Patocchi A (2014) Fine-mapping of the apple scab
resistance locus Rvi12 (Vb) derived from ‘Hansen’sbaccata#2
’.Mol
Breeding 34:2119–2129
Pagliarani G, Dapena E, Miñarro M, Denancé, Lespinasse Y, Rat-Morris,
Troggio, Durel CE, Tartarini S (2016) Fine mapping of the rosy
apple aphid resistance Dp-fl locus on Linkage Group 8 of the apple
cultivar ‘Florina’. Tree Genet Genom
Paris R, Cova V, Pagliarani G, Tartarini S, Komjanc M, Sansavini
S (2009) Expression profiling in HcrVf2-transformed apple
plants in response to Venturia inaequalis.TreeGenet
Genomes 5:81–91
Patocchi A, Vinatzer BA, Gianfranceschi L, Tartarini S, Zhang HB,
SansaviniS, Gessler C (1999) Construction of a 550 kb BAC contig
spanning the genomic region containing the apple scab resistance
gene Vf. Mol Gen Genet 262:884–891
Thomson MJ (2014) High-throughput SNP genotyping to accelerate crop
improvement. Plant Breed Biotechnol 2:195–212
Uitdewilligen JG, Wolters AMA, Bjorn B, Borm TJ, Visser RG, van Eck
HJ (2013) A next-generation sequencing method for genotyping-by-
sequencing of highly heterozygous autotetraploid potato. PlosOne
8(5), e62355. doi:10.1371/journal.pone.0062355
Varshney RK, Nayak SN, May GD, Jackson SA (2009) Next-generation
sequencing technologies and their implications for crop genetics and
breeding. Trends Biotechnol 27:522–530
VinatzerBA,PatocchiA,GianfranceschiL,TartariniS,Zhang
HB, Gessler C, Sansavini S (2001) Apple contains receptor-
like genes homologous to the Cladosporium fulvum resis-
tance gene family of tomato with a cluster of genes coseg-
regating with Vf apple scab resistance. Mol Plant Microbe
Interact 14:508–515
Vinatzer BA, Patocchi A, Tartarini S, Gianfranceschi L, Sansavini S,
Gessler C (2004) Isolation of two microsatellite markers from
BAC clones of the Vf scab resistance region and molecular
35 Page 20 of 21 Tree Genetics & Genomes (2016) 12:35
characterization of scab-resistant accessions in Malus germplasm.
Plant Breed 123:321–326
Voorrips RE, Bink MC, van de Weg WE (2012) Pedimap: software for
the visualization of genetic and phenotypic data in pedigrees. J
Hered. doi:10.1093/jhered/ess060
Zhu YM, Barritt BH (2008) Md-ACS1 and Md-ACO1 genotyping of
apple (Malus xdomestica Borkh.) breeding parents and suitability
for marker-assisted selection. Tree Genet Genomes 4:555–562
Data archiving statement
Sequences used to develop the KASP
TM
assays are released into the
NCBI dbSNP database. Accession numbers are indicated in Table 1.
Tree Genetics & Genomes (2016) 12:35 Page 21 of 21 35
A preview of this full-text is provided by Springer Nature.
Content available from Tree Genetics & Genomes
This content is subject to copyright. Terms and conditions apply.
