Construction of an integrated high density simple sequence repeat linkage map in cultivated strawberry (Fragaria × ananassa) and its applicability.

The cultivated strawberry (Fragaria × ananassa) is an octoploid (2n = 8x = 56) of the Rosaceae family whose genomic architecture is still controversial. Several recent studies support the AAA'A'BBB'B' model, but its complexity has hindered genetic and genomic analysis of this important crop. To overcome this difficulty and to assist genome-wide analysis of F. × ananassa, we constructed an integrated linkage map by organizing a total of 4474 of simple sequence repeat (SSR) markers collected from published Fragaria sequences, including 3746 SSR markers [Fragaria vesca expressed sequence tag (EST)-derived SSR markers] derived from F. vesca ESTs, 603 markers (F. × ananassa EST-derived SSR markers) from F. × ananassa ESTs, and 125 markers (F. × ananassa transcriptome-derived SSR markers) from F. × ananassa transcripts. Along with the previously published SSR markers, these markers were mapped onto five parent-specific linkage maps derived from three mapping populations, which were then assembled into an integrated linkage map. The constructed map consists of 1856 loci in 28 linkage groups (LGs) that total 2364.1 cM in length. Macrosynteny at the chromosome level was observed between the LGs of F. × ananassa and the genome of F. vesca. Variety distinction on 129 F. × ananassa lines was demonstrated using 45 selected SSR markers.

Introduction

The cultivated strawberry (Fragaria × ananassa) is one of the most popular and globally consumed fruit crops. It is cultivated in various regions throughout the world; in 2010, 39% of the cultivated strawberry crop was produced in North and South America, followed by Europe (33%), Asia (18%), and Africa (9%).[1] Because of the economic importance of this species, breeding programmes of cultivated strawberry are underway in many countries. To date, more than 3000 varieties bred in 41 countries have been registered in the International Union for the Protection and New Varieties of Plants variety database.[2] However, despite intensive use of F. × ananassa in the industry, the progress of genetic and genomic research of this crop has lagged behind that of many other economically important plant species, because of its complex genome structure.

Fragaria × ananassa is an octoploid (2n = 8x = 56) species that originated from a natural hybridization between Fragaria virginiana and Fragaria chiloensis.[3] The genome composition of F. × ananassa was initially proposed as AABBBBCC[4] or AAA′A′BBBB,[5] based on results of cytological studies. Later, Bringhurst[6] proposed the AAA′A′BBB′B′ model in light of cytological and genetic evidence. Unlike the first two models, the last model suggests disomic inheritance in entire chromosomes. Although the genome composition of F. × ananassa has not yet been confirmed, the AAA′A′BBB′B′ model has been supported by recent molecular genetic studies.[7-11] The genus Fragaria belongs to the Rosaceae family. Unlike other genera in the Rosaceae family, Fragaria comprises a limited number of species (approximately 22).[12] Several species have been nominated as candidate diploid ancestors such as Fragaria vesca, Fragaria iinumae, and Fragaria daltoniana.[13,14] Of these candidates, F. vesca is considered to be the most likely diploid ancestor, and it serves as a model species of F. × ananassa.[14] Therefore, genomic and genetic studies in F. vesca have been performed prior to those of F. × ananassa.

In molecular genetics studies, several types of F. vesca sequence-derived DNA markers, such as simple sequence repeat (SSR) and sequence-characterized amplified region (SCAR) markers, have been developed since 2003[15-19] (in this article, ‘DNA marker’ is defined as ‘unique DNA sequence(s) identified by specific primers’). Linkage maps were then constructed with a population derived from an interspecific cross between F. vesca and Fragaria nubicola.[20-23] In genomic studies, bacterial artificial chromosome and fosmid libraries were constructed to investigate features of genome sequences of F. vesca.[24,25] Subsequently, a whole genome sequence for F. vesca was published in 2011, and the results have greatly contributed to advances in genomic and genetic analysis of the genus Fragaria.[26] In parallel, comparative genomics studies within the Rosaceae family have also been performed. Cabrera et al.[27] developed a total of 1039 Rosaceae-conserved orthologous set (RosCOS) markers. Primer pairs of the RosCOS markers were designed on intron-flanking regions of orthologous genes commonly conserved among the genera Malus, Fragaria, and Prunus and mapped onto a Prunus reference linkage map. After the genome sequence in F. vesca was published,[26] whole genome comparisons between Fragaria, Prunus, and Malus were performed, and major orthologous regions were identified across the genus.[28]

In F. × ananassa, the first linkage map was constructed with 235 or 285 amplified fragment length polymorphism (AFLP) markers on 28 or 30 linkage groups (LGs) based on a 2-way pseudo-testcross strategy.[29] Several linkage maps were subsequently constructed with AFLP, SCAR, sequence-tagged site, random amplified polymorphic DNA, and SSR markers.[8-11,30] All the previously published linkage maps were generated based on single F1 mapping populations, and integrated linkage maps were developed in each mapping population in all four studies.[8,10,11,30] Of the previously developed linkage maps, the densest map was constructed by Sargent et al.[10]; a total of 549 loci were mapped onto 28 LGs, with a total length of 2140.3 cM. The density of the map was greatly enhanced over previous maps, but several unsaturated LGs were still observed such as the unintegrated LG pair of FG2DA and FG2DB and LG6B that contained large gaps (36.7 cM). These results suggest the need to develop a denser linkage map in F. × ananassa to reveal the complex genome structure of this species.

In this study, we performed SSR marker development and constructed an integrated high density linkage map to accelerate the advancement of genomic and genetic studies in F. × ananassa. Three types of SSR markers were developed, namely F. vesca expressed sequence tag (EST)-derived SSR markers (FVES markers), F. × ananassa EST-derived SSR markers (FAES markers), and F. × ananassa transcriptome-derived SSR markers (FATS markers), using public genome sequence data. The markers were mapped onto five parent-specific linkage maps, along with previously published DNA markers, and the five parent-specific maps were then integrated into one map. The applicability of the integrated map and the markers developed in this study were also demonstrated by comparative analysis of F. × ananassa and F. vesca and by variety distinction on 129 F. × ananassa lines with selected markers. The markers and integrated linkage map described in this study are valuable resources for future studies that will help to elucidate the genome structure and evolutionary process in F. × ananassa and whole genome sequencing, genetic mapping, and molecular breeding of this species.

Materials and methods

Plant material

An integrated linkage map was constructed using three mapping populations originating from a total of five parental lines. ‘02–19’ × ‘Sachinoka’ is an F1 mapping population of 188 individuals derived from a one-way pseudo-testcross. The female parent ‘02–19’ is a breeding line developed in Chiba Prefectural Agriculture and Forestry Research Center, which is resistant to powdery mildew and Fusarium wilt. The male parent ‘Sachinoka’ is a Japanese variety bred in the National Agricultural Research Center for Kyusyu and Okinawa Region, which was derived from a cross between ‘Toyonoka’ and ‘Aiberry’. ‘Kaorino’ × ‘Akihime’ consisted of 140 F1 individuals derived from a one-way pseudo-testcross between the parental lines. ‘Kaorino’ and ‘Akihime’ were bred in Mie Prefecture Agricultural Research Institute and by Mr Kazuhiro Ogiwara in Japan, respectively. One of the ancestor lines of ‘Kaorino’ is ‘Akihime’. ‘Kaorino’ exhibits crown rot resistance, whereas ‘Akihime’ is susceptible to the disease. The ‘0212921’ mapping population consisted of 169 S1 individuals of ‘0212921’. ‘0212921’ was generated from a cross between two breeding lines developed in Mie Prefecture Agricultural Research Institute, and the male parent of ‘0212921’ was identical to the female parent of ‘Kaorino’.

Development of EST-derived SSR markers

SSR regions were identified from the F. vesca and F. × ananassa ESTs registered in public databases, namely NCBI's nucleotide database (http://www.ncbi.nih.gov). The numbers of ESTs that were investigated include 45 739 and 6117 in F. vesca and F. × ananassa, respectively. SSRs longer than 14 bases, which contained all possible combinations of dinucleotide (NN), trinucleotide (NNN) and tetranucleotide (NNNN) repeats, were identified using the FINDPATTERNS module in the GCG software package (Accelrys Inc., USA). Oligonucleotides for polymerase chain reaction (PCR) primers were designed based on the flanking regions of the identified SSRs using the Primer3 program[31] in such a way that the amplified products ranged from 90 to 300 bp in length. Markers corresponding to previously published SSR markers in the Fragaria spp and RosCOS of markers[8,10,15,17-23,27,32-39] were identified based on primer sequences and were excluded from the collection of FVES and FAES markers.

Development of transcriptome-derived SSR markers

A total of 1 188 226 F. × ananassa transcript sequence reads registered in NCBI's Sequence Read Achive (SRA, Accession number: SRX16008) were used for identification of the SSR regions. All reads were generated using the Roche 454 sequencing system (Roche, Basel, Switzerland). The MIRA 3.2.0 program was used for assembling non-redundant contigs.[40] Methods for identification of SSR regions and primer design were the same as those used for FVES and FAES marker development, except that penta- and hexanucleotide repeats were also identified and used for primer design. After designing the primer sequences, the redundancy between the FVES, FAES, and published markers described in the above section was confirmed. The newly developed markers were designated FATS markers.

Polymorphic analysis of the DNA markers

DNA was extracted from the young leaves of plants using a DNeasy Plant Mini Kit (Qiagen Inc., Germany). DNA quantification and quality checks were performed using an ND1000 NanoDrop spectrophotometer (Nanodrop Technologies, DE, USA) and 0.8% agarose gels. In addition to the FVES, FAES, and FATS markers, a total of 1114 primer pairs of previously published SSR markers developed from Fragaria sequences and RosCOS markers were used for polymorphic analysis of the ‘02–19’ × ‘Sachinoka’ mapping population (Supplementary Table S1).[8,15,17-21,23,27,32-38] Polymorphic analysis of the other two populations was performed without using published markers. PCR was performed in a 5-µl reaction volume using 0.6 ng of genomic DNA in 1X PCR buffer (Bioline, UK), 3 mM MgCl2, 0.08 U of BIOTAQ DNA polymerase (Bioline, UK), 0.8 mM dNTPs, and 0.4 µM of each primer. A modified touchdown PCR protocol was followed as described by Sato et al.[41] The PCR products were separated by 10% polyacrylamide gel electrophoresis in tris-borate-ethylenediaminetetraacetic acid (TBE) buffer according to the standard protocol or with an ABI 3730xl fluorescent fragment analyzer (Applied Biosystems, USA), according to the polymorphic fragment sizes of the PCR amplicons. In the former case, the data were analysed using the Polyans software (http://www.polyans.kazusa.or.jp), whereas in the latter case, polymorphisms were investigated using the GeneMapper software (Applied Biosystems, USA).

Linkage analysis

In this study, it was assumed that F. × ananassa is an allo-octoploid species and that polymorphic loci in entire chromosomes showed disomic inheritance. Therefore, linkage analysis was performed using the methodology employed with diploid and outcrossing species. The segregated data of the two pseudo-testcross mapping populations were categorized into two parent-specific data sets by comparing the sizes of polymorphic bands of the parents and progeny. The segregation data were rescored using the ‘HAP1’ or ‘F2’ population type codes employed in JoinMap analysis.[42] ‘HAP1’ codes were used for the four parental lines of the ‘02–19’ × ‘Sachinoka’ and ‘Kaorino’ × ‘Akihime’, whereas ‘F2’ codes were employed for the parents of the S1 population ‘0212921’. As a result, a total of five parent-specific data sets were generated. The segregation data from each parent-specific data set were then classified into multiple LGs using the colour map method[43] that employed a comparison of graphical genotypes of the segregation data. During the process of colour mapping, reciprocal genotypes were converted to coupling genotypes. The robustness of the data sets for each LG was then confirmed using the Grouping Module of the JoinMap program, version 4, with a logarithm of odds (LOD) threshold of 10. The homeologous LGs within each parent-specific map were assumed based on corresponding positions on the F. vesca genome, which were predicted by comparative analysis (described below).

The locus orders in each parent-specific map were calculated using the Regression Mapping Module of JoinMap. ‘0212921’-specific data were handled as an ‘F2’ population type, whereas other parent-specific data sets were calculated as ‘HAP1’ population types. The following parameters were used for the calculation: Kosambi's mapping function, LOD ≥ 1.0, REC frequency ≤0.4, goodness of fit jump threshold for removal of loci = 5.0, number of added loci after which a ripple is performed = 1, and third round = yes. Each LG in parent-specific maps was named according to the following rule: numbers after LG (1–7) showed corresponding chromosome numbers of F. vesca predicted by comparative analysis (described below). The capital letters A–D distinguished four LGs in each homeologous group (HG), which corresponded to LGs in an integrated map. The A–D suffixes corresponded to the lengths of the LGs on the integrated map. When multiple LGs were integrated into a single LG of the integrated map, they were numbered using capital letters with underbars (e.g. LG1B_1 and LG1B_2). For unintegrated LGs, X or Y was used as a suffix.

For construction of an integrated linkage map, corresponding LGs between the parent-specific maps were assumed by identification of commonly mapped markers on each LG. Prior to integration, genotypes of dominant loci in each parent-specific map were imputed to co-dominant genotypes, according to the flanking genotypes of co-dominant loci. That is, alleles of dominant loci were converted to co-dominant alleles, according to the flanking genotypes of co-dominant loci. Parent-specific locus genotype data sets were then integrated into one dataset in each LG using the Combine Groups for Mapping Integration Module, followed by locus ordering by the Regression Mapping Module of JoinMap. The parameters used for the mapping module of an integrated map were the same as those used for parent-specific mapping. After construction of an integrated map, the mapped loci were classified into two groups, i.e. loci generated from single-locus diagnostic (SLD) and multi-loci diagnostic (MLD) markers. SLD markers were defined as markers that detected single segregation bands mapped onto single positions on an integrated map, and not amplified other loci, such as monomorphic loci, whereas MLD markers were defined as markers that amplified more than one locus, including monomorphic loci.

Comparative mapping

Syntenic regions between the genomes of F. × ananassa and F. vesca were detected by identifying the conservation of the relative locations of genes and genomic regions. The source of the genome sequences of F. vesca was obtained from NCBI [Accession numbers: CM001053.1-CM001059.1 (LG1-7), GG775183.1- GG775301.1 (unplaced)]. The EST sequences adjacent to the mapped markers on the integrated F. × ananassa map were compared with the gene sequences in the reference genomes using the BLASTX program with a cutoff E-value ≤ 1e-10. The syntenic regions defined by the top hits of the homology search were plotted using the Cicros program (http://circos.ca).

Distinguishing the varieties of 129 F. × ananassa lines using selected EST-SSR markers

In order to distinguish the varieties of 129 F. × ananassa lines, a total of 22 F. × ananassa varieties and breeding lines, most of which were bred in Japan, were used in pre-polymorphic analysis, including the following: ‘Sachinoka’, ‘Fusanoka’, ‘Asuka Wave’, ‘Kaorino’, ‘Akihime’, ‘Miyoshi’, ‘Dover’, ‘Strawberry Parental Line Nou-2’, ‘Karenberry’, ‘Ohkimi’, ‘Sanchiigo’, ‘Toyonoka’, ‘Nyohou’, Tochiotome’, ‘Nou-Hime’, ‘Asuka ruby’, ‘Sagahonoka’, ‘Beni Hoppe’, ‘Yayoihime’, ‘Sanukihime’, ‘02-19’, and ‘0212921’. PCR was performed with all the FVES and FAES markers developed in this study. The PCR protocol was same as that used for polymorphic analysis of the DNA markers (see above). The PCR products were separated by 10% polyacrylamide gel electrophoresis in TBE buffer using the standard protocol. SSR markers showing solid and polymorphic amplification were selected, and the robustness of PCR, i.e. the repeatability of results and the absence of noise peaks, was investigated for the selected markers using a fragment analyzer (ABI 3730xl, Applied Biosystems, USA) for 22 F. × ananassa lines. PCR was performed in a 5-µl reaction volume using 1 ng of genomic DNA in 1X PCR buffer II (Applied Biosystems, USA), 2.5 mM MgCl2, 0.125 U of AmpliTaq Gold DNA polymerase (Applied Biosystems, USA), 0.2 mM dNTPs, and 0.4 µM of each primer. The thermal cycling conditions were as follows: 7 min denaturation at 95°C; 30 cycles of 30 s denaturation at 95°C, 30 s annealing at 57°C, and 1 min extension at 72°C, with a final 10 min extension at 72°C. A total of 100 markers showing high repeatability were screened, and the robustness of PCR was again confirmed for 129 F. × ananassa lines, 119 of which were previously investigated polymorphisms identified using CAPS markers[44] (Supplementary Table S2). The 45 best SSR markers were thereby selected to distinguish the varieties of 129 strawberry lines. The allelic data were converted into a binary matrix using the scores 1 or 0 for the presence or absence of the peak. The expected heterozygosity (HZ) of each identified peak was calculated using the following formula:

where P and P are the frequency of presence and absence of the ith peak. In F. × ananassa, SSR markers often identify multi loci due to polyploidy, and it is often difficult to investigate the exact number of loci identified in each marker in an unrelated population. Therefore, the mean HZ value of identified peaks generated by a marker was substituted for the HZ of each marker. The allelic binary data were also analysed using GGT 2.0[45] for the investigation of genetic similarity between the varieties using a Jaccard similarity coefficient. A unweighted pair group method (UPGMA) with arithmetic average dendrogram was constructed using MEGA version 5.05.[46]

Results

Development of EST-SSR markers

A total of 2748 and 324 SSRs were identified in non-redundant 15 203 F. vesca and 1029 F. × ananassa ESTs, respectively. Of the identified SSRs, 562 primers were designed based on the flanking regions of 453 F. vesca- and 109 F. × ananassa-derived EST-SSRs and designated FVES and FAES markers. To increase the number of EST-SSR markers, additional primer pairs were designed that allowed either single-base mismatches (779 and 107 primer pairs for FVES and FAES markers, respectively) or two-base mismatches (2514 and 387 primer pairs for FVES and FAES markers, respectively) in the SSR regions. Of the markers that were generated, those corresponding to previously published markers were excluded.[8,10,15,17-23,27,32-39] As a result, a total of 3746 FVES and 603 FAES markers were developed. Design details of the FVES and FAES markers, including primer sequences, corresponding SSR motifs, expected product sizes, and GenBank IDs of the template EST sequences, are listed in Supplementary Tables S3 and S4 and on the web at http://marker.kazusa.or.jp/strawberry/.

Of the FVES and FAES markers that were developed, 2975 (79.4%) and 431 (71.5%) were trinucleotide repeats while 456 (12.2%) and 117 (19.4%) were dinucleotide repeats, and 315 (8.4%) and 55 (9.1%) were tetranucleotide repeats, respectively (Table 1). In the FVES markers, the poly (AAG)n motif was most abundant in the trinucleotide repeats (755 SSRs, 20.1%), followed by poly (GGA)n (492 SSRs, 13.1%), poly (ATC)n (352 SSRs, 9.4%), poly (AG)n (347 SSRs, 9.3%), and poly (AGC)n (321 SSRs, 8.6%). Like the FVES markers, the poly (AAG)n motif was the most frequently observed motif in the FAES markers (132 SSRs, 21.9%). This was followed by poly (AG)n (81 SSRs, 13.4%), poly (ATC)n (68 SSRs, 11.3%), and poly (AGC)n (48 SSRs, 8.0%). Among the tetranucleotide repeats, AT-rich motifs, namely poly (AAAG)n, and poly (AAAC)n, were more frequently observed than other motifs in both the FVES and FAES markers.

Table 1.

Numbers of SSR motifs in the FVES, FAES, and FATS markers

Motif	FVES				FAES				FATS
	Designed (%)		Mapped (%)		Designed (%)		Mapped (%)		Designed (%)		Mapped (%)
Dinucleotide
AG	347	(9.3)	120	(11.0)	81	(13.4)	34	(17.7)	10	(8.0)	4	(10.3)
AT	60	(1.6)	21	(1.9)	22	(3.6)	12	(6.3)	0	(0.0)	0	(0.0)
AC	49	(1.3)	18	(1.7)	14	(2.3)	5	(2.6)	1	(0.8)	0	(0.0)
TC	0	(0.0)	0	(0.0)	0	(0.0)	0	(0.0)	7	(5.6)	4	(10.3)
Others	0	(0.0)	0	(0.0)	0	(0.0)	0	(0.0)	7a	(5.6)	0	(0.0)
Subtotal	456	(12.2)	159	(14.6)	117	(19.4)	51	(26.6)	25	(20.0)	8	(20.5)
Trinucleotide
AAG	755	(20.1)	210	(19.3)	132	(21.9)	41	(21.4)	5	(4.0)	1	(2.6)
GGA	492	(13.1)	138	(12.7)	45	(7.5)	6	(3.1)	4	(3.2)	3	(7.7)
ATC	352	(9.4)	104	(9.6)	68	(11.3)	16	(8.3)	2	(1.6)	1	(2.6)
AGC	321	(8.6)	92	(8.4)	48	(8.0)	20	(10.4)	3	(2.4)	2	(5.1)
GGC	291	(7.8)	77	(7.1)	27	(4.5)	11	(5.7)	1	(0.8)	0	(0.0)
GGT	268	(7.2)	78	(7.2)	40	(6.6)	7	(3.6)	1	(0.8)	0	(0.0)
AAC	207	(5.5)	58	(5.3)	34	(5.6)	11	(5.7)	0	(0.0)	0	(0.0)
ACG	168	(4.5)	50	(4.6)	16	(2.7)	5	(2.6)	0	(0.0)	0	(0.0)
ACT	62	(1.7)	8	(0.7)	5	(0.8)	0	(0.0)	0	(0.0)	0	(0.0)
AAT	59	(1.6)	17	(1.6)	16	(2.7)	6	(3.1)	0	(0.0)	0	(0.0)
CTT	0	(0.0)	0	(0.0)	0	(0.0)	0	(0.0)	4	(3.2)	1	(2.6)
Others	0	(0.0)	0	(0.0)	0	(0.0)	0	(0.0)	36a	(28.8)	12	(30.8)
Subtotal	2975	(79.4)	832	(76.4)	431	(71.5)	123	(64.1)	56	(44.8)	20	(51.3)
Tetranucleotide
AAAG	100	(2.7)	37	(3.4)	15	(2.5)	4	(2.1)	0	(0.0)	0	(0.0)
AAAC	54	(1.4)	14	(1.3)	7	(1.2)	1	(0.5)	0	(0.0)	0	(0.0)
AAAT	43	(1.1)	11	(1.0)	14	(2.3)	6	(3.1)	0	(0.0)	0	(0.0)
GGGA	26	(0.7)	8	(0.7)	3	(0.5)	2	(1.0)	0	(0.0)	0	(0.0)
AAGC	21	(0.6)	6	(0.6)	2	(0.3)	1	(0.5)	0	(0.0)	0	(0.0)
AATC	17	(0.5)	4	(0.4)	5	(0.8)	1	(0.5)	0	(0.0)	0	(0.0)
AATG	15	(0.4)	7	(0.6)	1	(0.2)	1	(0.5)	0	(0.0)	0	(0.0)
Others	39	(1.0)	11	(1.0)	8	(1.3)	2	(1.0)	0	(0.0)	0	(0.0)
Subtotal	315	(8.4)	98	(9.0)	55	(9.1)	18	(9.4)	0	(0.0)	0	(0.0)
Pentanucleotide
Subtotal	—	—	—	—	—	—	—	—	7a	(5.6)	1	(2.6)
Hexanucleotide
Subtotal	—	—	—	—	—	—	—	—	37a	(29.6)	10	(25.6)
Total	3746	(100)	1089	(100)	603	(100)	192	(100)	125	(100)	39	(100)

aThe numbers of types of observed ‘other’ SSR motifs in di- and trinucleotide repeats and all penta- and hexa- nucleotide repeats were 4, 24, 6, and 37, respectively.

Development of transcriptome-SSR markers

A total of 1 188 226 SRA sequences were assembled into 80 430 contigs by the MIRA 3.2.0 program.[38] On these contigs, 34 993 SSRs were identified. Primers that targeted the flanking regions of 129 of the 34 993 SSRs were designed and synthesized. All the 129 SSRs were located on non-redundant contigs, and a total of 125 primers, which did not overlap with the FVES, FAES, and previously published markers,,[8,10,15,17-23,27,32-39] were designed as FATS markers. The primer sequences of the FATS markers, along with the corresponding SSR motifs, product sizes, primer sequences, and template contigs, are provided on the web at http://marker.kazusa.or.jp/strawberry/ and in Supplementary Table S5. Of the 125 FATS markers, trinucleotide repeats were the most frequently observed (56 SSRs, 44.8%), followed by hexanucleotide repeats (37 SSRs, 29.6%), and dinucleotide repeats (25 SSRs, 20.0%, Table 1). No tetranucleotide repeats were observed, whereas seven pentanucleotide repeats were identified. Of the SSR motifs, the poly (AG)n motif was the most abundant (10 SSRs, 8.0%), followed by poly (TC)n (7 SSRs, 5.6%), and poly (AAG)n (5 SSRs, 4.0%). In the penta- and hexanucleotide repeats, each single SSR was identified for all the observed motifs except poly (GCTGT)n (two SSRs, data not shown).

Construction of parent-specific linkage maps

The ‘0212921’ S1 mapping population

A total of 4474 primer pairs of SSR markers, including 3746 FVES, 603 FAES, and 125 FATS, were investigated with 8 randomly chosen S1 individuals of the ‘0212921’ mapping population. Polymorphisms were observed on 605 FVES, 135 FAES, and 29 FATS markers, and segregation analysis was performed for 169 S1 individuals with 769 primer pairs. A total of 881 segregation locus genotypes were generated from the 769 primer pairs, and 822 of the 881 loci were mapped onto 34 LGs (Supplementary Tables S6 and S7). The length of each LG ranged from 1.5 cM (LG4B) to 94.5 cM (LG1A), representing a total length of 1508.3 cM. The mean locus density and segregation distortion (P < 0.05) were 1.83 cM locus−1 and 22.4%, respectively, ranging from 0.22 cM locus−1 (LG6D) to 4.26 cM locus−1 (LG2A_2) and from 0.0% (LG4B) to 85.7% (LG5A_1), respectively.

The ‘02–19’ × ‘Sachinoka’ mapping population

Polymorphic analysis was performed with a total of 5588 primer pairs of SSR markers, including 4474 SSR markers developed in this study and 1114 previously published markers (Supplementary Table S1). A total of 1299 markers, i.e. 853 FVES, 131 FAES, 37 FATS, and 278 published markers, showed polymorphisms between the parental lines. By performing segregation analysis of the 188 individuals of the mapping population, 1078 polymorphic loci were generated from 881 of the 1299 primer pairs that were tested. Of the 1078 segregation loci, 260 showed biparental polymorphisms while the other 818 were parent specific. A total of 575 and 556 loci were mapped onto ‘02–19’ and ‘Sachinoka’ specific maps, respectively (Supplementary Tables S6 and S7). In the ‘02–19’ specific map, 32 LGs were constructed on 1668.9 cM, with lengths ranging from 2.0 cM (LG2X) to 129.2 cM (LG1A). The mean locus density and segregation distortion (P < 0.05) were 2.90 cM locus−1 and 30.2%, respectively, ranging from 0.34 cM locus−1 (LG5A_2) to 11.63 cM locus−1 (LG6C) and from 0.0% (LG3B_2, 5A_2 and 7D_2) to 100% (LG2X and 4B), respectively. In the ‘Sachinoka’ map, 34 LGs were developed, totalling 2166.6 cM in length, with a mean locus density of 3.90 cM locus−1. The length and locus density of each LG ranged from 1.6 cM (LG7C_2) to 120.0 cM (LG7A) and from 0.80 cM locus−1 (LG7C_2) to 16.08 cM locus−1 (LG3X), respectively. Segregation distortion (P < 0.05) of each LG ranged from 0.0% (LG7C_2, 7D_1 and 7X) to 85.7% (LG2X), with a mean value of 25.4%.

The ‘Kaorino’ × ‘Akihime’ mapping population

A total of 4474 primer pairs of SSR markers developed in this study were used for polymorphic analysis with the parental lines. Polymorphisms were observed in 438 FVES, 95 FAES, and 33 FATS markers, and segregation analysis was performed with the 140 mapping plants, for a total of 566 primer pairs. A total of 612 segregation locus genotypes were generated from 537 of the 566 primer pairs that were tested. Of the 612 segregation loci, 140 showed biparental polymorphisms while the other 472 were parent specific. A total of 294 and 318 loci were mapped onto ‘Kaorino’ and ‘Akihime’ specific maps, respectively (Supplementary Tables S6 and S7). The ‘Kaorino’ map consisted of 32 LGs, representing a total length of 1103.4 cM. The mean locus density and segregation distortion (P < 0.05) were 3.75 cM locus−1 and 24.8%, respectively, ranging from 0.45 cM locus−1 (LG1B_2) to 7.61 cM locus−1 (LG1A) and from 0.0% (LG1B_2, 1X, 2A, 2D, 4X, 5A_1, 5X, 6A_2, 6A_3 and 7D) to 100% (LG4D_2), respectively. In the ‘Akihime’ map, 33 LGs were developed, with a total length of 951.4 cM and a mean locus density of 2.99 cM locus−1. The length and locus density of each LG ranged from 2.2 cM (LG5X) to 77.7 cM (LG7A) and from 0.55 cM locus−1 (LG5X) to 12.72 cM locus−1 (LG6A_4), respectively. Segregation distortion (P < 0.05) of each LG ranged from 0.0% (LG3X, 4A_2, 4C, 5A, 5C, 5X, 6A_1, 6B_1, 6B_2, and 7D_2) to 100% (LG1A_1), with a 14.8% mean value.

Construction of an integrated linkage map

Prior to integration, the subsets of parent-specific LGs to be integrated were determined by the numbers of commonly mapped markers across parent-specific maps, that is, pairs of LGs showing the largest number of commonly mapped markers were assembled into the same subset. When one LG showed the same numbers of commonly mapped markers in different pairs, the LG was excluded from the integration. Locus genotype data of each subset of LG were integrated using the Combine Groups for Mapping Integration Module in JoinMap, and the locus orders of an integrated map were then calculated. When grouped LG subsets were inadequate, the segregation loci of incorrect LGs were excluded during the process of locus ordering. When a LG subset was misassembled with incorrect pairs of parent-specific LGs, the loci of the mis-integrated LG overlapped with loci onto other integrated LGs which correctly paired LGs were integrated. Therefore, the correct assembly of a LG subset was determined by confirming the number of overlapping loci on each integrated LG using the Mapping module in JoinMap.

The integrated linkage map consisted of 1856 loci on 28 LGs, totalling 2364.1 cM in length (Table 2, Supplementary Table S7 and Fig. S1). The length of each LG varied from 34.2 (LG1D) to 123.7 cM (LG1A). The mean locus density was 1.27 cM locus−1, ranging from 0.68 cM locus−1 (LG6B) to 5.08 cM locus−1 (LG4B). The largest gap was 27.3 cM, between FVES1598_7a and FVES1351_7a on LG7A, followed by a 23.4 cM gap between FAES0326_4b and FAES3462_4b on LG4B, and a 21.6 cM gap between FVES0576_4c and FAES0045 on LG4C. The number of integrated LGs ranged from 1 (LG1D) to 16 (LG6A) parent-specific LGs (Table 2). The number of unintegrated parent-specific LGs, which totalled 17, ranged from 1 (‘0212921’) to 5 (‘Akihime’) in each parental map. The ratios of mapped FVES, FAES, and FATS loci were 77.5, 15, and 3.4%, respectively (Supplementary Fig. S2). The ratio of mapped FVES markers in each LG ranged from 62.5 to 93.3%.

Table 2.

Number of mapped loci, length, locus density in an integrated map, and numbers of integrated LGs and loci of parental specific maps

LG	Interspecific map				Parental specific maps
					‘0212921’		‘02–19’		‘Sachinoka’		‘Kaorino’		‘Akihime’
	Number of mapped loci	Length (cM)	Locus density (cM)	Single locia (%)	LGsb	Locic	LGsb	Locic	LGsb	Locic	LGsb	Locic	LGsb	Locic
1A	91	123.7	1.36	16 (17.5)	1	48	1	40	1	13	1	7	2	10
1B	81	80.7	1.00	25 (30.8)	1	25	1	41	2	42	2	14	0	0
1C	37	77.2	2.09	11 (29.7)	1	24	1	9	1	13	0	0	0	0
1D	24	34.2	1.43	5 (20.8)	1	24	0	0	0	0	0	0	0	0
2A	123	113.9	0.93	41 (33.3)	2	25	2	50	1	67	1	8	2	22
2B	71	95.1	1.34	15 (21.1)	1	42	2	20	1	4	1	11	1	15
2C	55	93.8	1.71	13 (23.6)	1	28	0	0	1	18	1	24	0	0
2D	64	81.1	1.27	19 (29.6)	1	37	1	28	1	7	1	8	0	0
3A	114	96.0	0.84	47 (41.2)	1	33	2	70	2	27	1	23	1	18
3B	89	82.7	0.93	19 (21.3)	1	40	2	37	1	24	0	0	1	11
3C	51	70.8	1.39	13 (25.4)	1	36	0	0	1	13	1	11	0	0
3D	38	65.8	1.73	5 (13.1)	1	30	0	0	0	0	1	9	1	3
4A	88	112.2	1.28	31 (35.2)	1	24	1	45	2	19	0	0	2	42
4B	18	91.4	5.08	2 (11.1)	1	3	1	5	1	13	0	0	0	0
4C	64	87.9	1.37	6 (9.37)	2	32	2	16	1	12	1	15	1	13
4D	45	53.0	1.18	7 (15.5)	1	27	0	0	1	13	2	14	1	5
5A	130	119.0	0.92	40 (30.7)	2	25	2	53	2	77	2	9	1	9
5B	57	79.3	1.39	13 (22.8)	1	35	0	0	0	0	1	26	2	15
5C	18	74.8	4.16	3 (16.6)	0	0	1	12	0	0	0	0	1	10
5D	23	46.0	2.00	5 (21.7)	1	10	0	0	0	0	1	15	0	0
6A	131	109.8	0.84	31 (23.6)	3	37	4	37	2	44	3	16	4	49
6B	129	87.9	0.68	32 (24.8)	2	86	1	29	2	31	2	15	2	9
6C	59	87.4	1.48	11 (18.6)	1	43	1	4	1	18	1	6	0	0
6D	49	76.4	1.56	10 (20.4)	1	11	0	0	1	15	2	24	2	23
7A	80	115.7	1.45	24 (30.0)	1	27	2	25	1	33	1	17	1	24
7B	37	72.4	1.96	11 (29.7)	1	30	0	0	0	0	1	5	1	9
7C	24	68.6	2.86	6 (25.0)	1	14	0	0	2	15	0	0	0	0
7D	66	67.3	1.02	20 (30.3)	1	21	2	35	2	14	1	3	2	14
Unintegrated	—	—	—	—	1	5	3	20	4	24	4	14	5	17
Total	1856	2364.1	1.27	481 (25.9)	34	822	32	576	34	556	32	294	33	318

aNumber of mapped loci generated from SLD markers. Numbers in parentheses show percentage of all the mapped loci.

bNumber of integrated LGs.

cNumber of integrated loci.

Of the mapped loci, 481 (25.9%) were generated from SLD markers, whereas the other 1375 were from MLD markers. The ratio of mapped loci generated from SLD markers in each LG varied from 9.4% (LG4C) to 41.2% (LG3A). These loci were mapped randomly onto many of the LGs while several clusters were observed on parts of LGs (LG1A, 1C, 2C, 4C, 4D, 6C, 6D, 7B, and 7C, Supplementary Fig. S1). Loci generated from MLD markers were classified into four types; multi loci mapped onto homoeologous LGs (Multi_H in Supplementary Table S7), non-homeologous LGs (Multi_NH), the same LGs (Multi_S), and single position (Multi_NM). For the Multi_NM loci, all the observed multiple bands were monomorphic except for mapped ones. Of the four types of multi loci, Multi_H was most frequently observed (684 loci), followed by Multi_NM (523), Multi_S (206), and Multi_NH (77). Of the 684 Multi_H loci, 16 and 99 were observed, corresponding to loci on non-homeologous LGs (Multi_H&NH) and the same LGs (Multi_H&S), respectively. The Multi_H loci were randomly distributed along the entire integrated linkage map, whereas Multi_S loci were not observed on several LGs, i.e. LG2B, 2D, 3C, 4D, 5A, 5B, 7B, and 7C (Supplementary Fig. S1). In each HG, the homeologous regions differed depending on the paired LGs (Supplementary Fig. S3). For example, most homeologous regions of LG2A and LG2C were observed on 20–40 cM while those of LG2B and LG2D were identified on 60–80 cM. In the case of HG1, homeologous regions were not observed between LG1C and LG1D. Such biases of homeologous pairs were identified across the genome.

Comparison with the genomes of a wild relative, F. vesca

Of the 3746 ESTs that corresponded to the designed FVES markers, 3743 showed significant similarities to the genome sequences of F. vesca, while 3 ESTs, corresponding to FVES1248, FVES2637, and FVES2807, did not (Supplementary Table S3). Of the 3743 ESTs, 3608 showed similarities to genome sequences placed on 7 chromosomes of F. vesca, and the other 135 were mapped onto unplaced genomic scaffolds. In F. × ananassa sequence-derived markers, significant similarities to the F. vesca genome sequences were observed in all 603 ESTs and 124 of the 125 transcript contigs from which FAES and FATS markers were designed, respectively (Supplementary Table S4 and S5). Of these, 593 and 120 sequences corresponding to FAES and FATS markers, respectively, were mapped onto F. vesca genome sequences placed on the 7 chromosomes, whereas the other 14 ESTs or transcript contigs were mapped onto unplaced scaffolds. Similarity searches were also performed for the previously published markers that were located on the integrated map. Of the 91 mapped previously published markers, 68 ESTs or genome sequence are available on the NCBI dbGSS database (http://www.ncbi.nlm.nih.gov/projects/dbGSS, Supplementary Table S1) from which the markers were designed. Of the 68 sequences, 62 showed significant similarities to the F. vesca genome sequences placed onto the 7 chromosomes.

A total of 1354 markers that generated 1783 mapped loci onto the integrated linkage map showed significant similarity with F. vesca genome sequences placed onto the 7 chromosomes. By considering the sequences with highest similarity scores to be putative orthologs, the map locations of the SSR markers and the corresponding F. vesca genome sequences were compared. As shown in Fig. 1, the alignment of homologous sequence pairs along each LG revealed an obvious syntenic relationship between corresponding HGs in F. × ananassa (Fa-HG) and chromosomes (Chr) in F. vesca (Fv-Chr). Most of the regions did not show synteny between non-corresponding Fa-HGs and Fv-Chr, except between Fa-HG3 and 25–30 Mb on Fv-Chr2. In some LGs in F. × ananassa, syntenic regions were observed on whole corresponding chromosomes in F. vesca, whereas segmental syntenic blocks were identified between pairs of the other Fa-LGs and Fv-Chr (Supplementary Fig. S3-1). For example, Fa-LG1A displayed syntenies against entire regions of Fv-Chr1, whereas segmental syntenic blocks were observed between 0 and 11 Mb in Fv-Chr1 and Fa-LG1B or Fa-LG1D, as well as between 8 and 23 Mb in Fv-Chr1 and Fa-LG1C.

Figure 1.

Graphical view of syntenic relationship between F. × ananassa and F. vesca. Red and blue bars show LGs of F. × ananassa and chromosomes of F. vesca, respectively. Syntenic regions between the two species are connected by coloured lines.

To select the best marker set to determine the varieties of the F. × ananassa lines, polymorphic analysis of the 22 cultivated strawberry lines was performed using primer pairs targeting 3746 FVES and 603 FAES markers. In the primer pairs of FVES markers, 2949 resulted in solid amplification, whereas of the remaining primer pairs, 431, 263, and 103 resulted in no amplification, multiple bands, and rare bands, respectively (Supplementary Table S3). In the FAES markers, solid amplification was observed for 460 primer pairs, whereas of the remaining primer pairs, 88, 52, and 3 resulted in no amplification, multiple, and rare bands, respectively (Supplementary Table S4). A total of 751 primer pairs, including 650 FVES and 101 FAES, showed solid and polymorphic amplification and were used for the subsequent investigation of PCR, i.e. the stability of PCR and detection of noise peaks, using a fragment analyzer for all 22 lines. A total of 100 of the 751 primer pairs showing high stability and few noise peaks were selected and used for polymorphic analysis with 129 strawberry lines to reconfirm the robustness of PCR in a more diversified collection. From this experiment, the 45 best markers, which showed high stability in PCR and few noise peaks, were selected. PCR was then performed twice more with the 129 strawberry lines to exclude genotyping errors (Supplementary Table S3).

The primer pairs of 45 selected SSR markers, including 4 FAES and 41 FVES markers, generated a total of 158 peaks in the 129 strawberry lines that were tested (http://vim.kazusa.or.jp/Strawberry/). The number of identified peaks per a primer pair ranged from two to seven, with a mean value of 3.51 (Supplementary Fig. S4). The HZ value of each identified peak was calculated based on a score of 1 or 0 for the presence and absence of the peak and ranged from 0 to 0.5. The mean HZ value of each marker ranged from 0.11 to 0.43, with an average value of 0.25. Similarity coefficients were used to examine the genetic relationships between the 129 strawberry lines. All possible genotypes showed similarity coefficients ranging from 0.00 to 0.41 (Supplementary Table S2). No genetic differences were identified with the 45 markers between ‘Nyoho’ and ‘Shinnyoho’ or between ‘Himatsuri’ and ‘Toyonoka’. The highest similarity coefficient, 0.41, was found between the UK variety ‘Serenata’ and the Japanese variety ‘Summer berry’.

Discussion

In this study, a total of 4474 SSR markers, including 3746 FVES, 603 FAES, and 125 FATS, were designed from public EST and transcript sequences of F. vesca and F. × ananassa. Of the 4474 SSR markers, 672 resulted in rare or no amplification, and the remaining 3802 SSR markers resulted in amplification. To our knowledge, 441 SSR markers have previously been published for the genus Fragaria.[8,10,15,17-23,27,32-39] The number of SSR markers that were developed in this study is approximately 8.6 times that of previously published markers. Recent advances in genome sequencing technology have enabled the large-scale development of single nucleotide polymorphism (SNP) markers in many of plant species. However, SNP discovery and genotyping are still difficult for polyploidy species because of the difficulty in distinguishing between homeologous allelic SNPs. Therefore, SSR markers are currently the most rapid and conclusive marker system for most polyploid species, including F. × ananassa. The numerous SSR markers developed in this study are an important resource for genetic and genomic studies of this species.

In this study, linkage analysis was performed using a methodology previously employed in diploid and outcrossing species with the assumption that F. × ananassa is an allo-octoploid species with an AAA′A′BBB′B′ genome structure. The large number of mapped loci that showed disomic inheritance supported this assumption.

In the ‘02–19’ × ‘Sachinoka’ and ‘Kaorino’ × ‘Akihime’ populations, polymorphic marker screening was only performed with parental lines that caused failure of screening polymorphic markers showing a segregation pattern. Meanwhile, eight randomly selected S1 individuals were used in polymorphic marker screening of the ‘0212921’ population, and all possible polymorphic markers were selected. Therefore, we considered that the large differences of the number of mapped loci in the parent-specific maps were due to the steps in polymorphic marker screening and the genetic distances of haplotypes within parents. All the previously published maps in F. × ananassa were derived from F1 mapping populations, and the ‘0212921’ specific map is the first parent-specific map derived from an S1 population. The density of the parent-specific map ‘0212921’ suggests that an S1 population is available for linkage map construction in F. × ananassa.

According to a previous study[10] and the integrated map developed in this study, the length of a saturated linkage map in F. × ananassa would exceed 2000 cM. The lengths of all the parent-specific maps were shorter than 2000 cM, except the ‘Sachinoka’ map that was 2166.6 cM. In addition, no parent-specific map had 28 LGs, i.e. the number of chromosomes in the haploid genome of F. × ananassa. Therefore, it was concluded that all parental specific maps were unsaturated.

The integrated linkage map comprises 1856 loci on 28 LGs. The total length of the map, 2364 cM, is slightly longer than the previously reported densest map, 2140 cM, by Sargent et al.[10] The mean locus density of our integrated map was 1.27 cM locus−1, which was three times denser than that of the map of Sargent et al. Moreover, the integrated map presented in this study is the first linkage map derived from multiple mapping populations of F. × ananassa, suggesting that the map reflects wider genetic diversity in the species. Several LGs of the integrated map seemed to be saturated, whereas some of others were not, such as LG1D, LG4B, LG5C, LG5D, LG7B, and LG7C. In addition, the parent-specific LGs were not evenly merged into the LGs in the integrated map. For example, LG6A in the integrated map consists of 16 parent-specific LGs, whereas LG1D was derived from a single parent-specific LG of the ‘0212921’-specific map. One of the causes for the uneven positions of the mapped loci and the number of merged parent-specific LGs may be the bias of heterozygous (polymorphic) genomic regions within each parental line. Moreover, the source of EST-SSR markers might affect the evenness of the mapping. Of the mapped loci on the integrated map, 82% were derived from F. vesca sequences, whereas 14 and 4% markers were generated from F. × ananassa and other species, respectively. Previous molecular and cytogenetic studies suggested that F. vesca and F. iinumae or F. daltoniana were candidate ancestral species of F. × ananassa.[13,14] This theory suggests that partial genome regions in F. × ananassa were derived from F. iinumae or F. daltoniana and may show non-homology with the genomes of F. vesca. In addition, the ratio of mapped FVES loci ranged from 62.5 to 93.3% (Supplementary Fig. S2). The large contribution of F. vesca sequence-derived SSR markers on the integrated map may affect the biased position of the mapped loci.

Although a total of 107 loci derived from previously published markers were mapped onto the integrated linkage map, we did not employ the corresponding names used in previously published maps. This was due to the insufficient number of commonly mapped markers across linkage maps, along with the large number of mapped multi loci. In this study, ‘A’ to ‘D’ suffixes were added to the LG names, which corresponded to the lengths of the LGs; A was used for the longest LG, whereas D was added to the shortest. Therefore, there are no biological relationships among the LGs designated with the same capital letter suffix. We request the replacing of the capital letters as to represent the genome structure in F. × ananassa, for example AAA′A′BBB′B′, after corresponding LGs to each genome, will be identified in future.

Loci generated from SLD markers should play an important role in the identification of specific LGs and chromosomes in F. × ananassa. In this study, we mapped both SLD and MLD markers, and 25.9% (481) of the mapped loci were derived from SLD markers. Indeed, the karyotypes of octoploid Fragaria species were already reported in F. chiloensis and F. virginiana,[47] as well as in F. × ananassa.[48] Using genomic region mapped single loci as probes for cytological analysis such as fluorescence in situ hybridization, corresponding karyotypes of F. × ananassa could be identified on each LG in the integrated map. In addition, the genome structure of F. × ananassa could be resolved by comparative mapping with candidate ancestral Fragaria species using markers detecting single loci.

While 25.9% of the mapped loci were generated from SLD markers, the remaining 74.1% (1375) were derived from MLD markers, comprising 569 Multi_H, 16 Multi_H&NH, 99 Multi H&S, 61 Multi_NH, 107 Multi_S, and 523 Multi_NM. The ratio of all Multi_H (including Multi_H&NS and H&S) to the sum of mapped loci (except Multi_NM) was 0.51 [(569 + 16 + 99)/(1856–523)]; that is approximately half of the mapped loci on the integrated map were Multi_H. The Multi_H loci were mapped along the entire genomes. However, corresponding homeologous positions differed depending on the paired LGs. The uneven homeologous regions between each pair of LGs might represent a feature of the genome composition of F. × ananassa.

Comparative mapping studies between F. × ananassa and F. vesca have been reported.[8-11] According to previous results, most of the genomic regions are conserved between the two species, with the rare exception of chromosome rearrangement on HG1, HG3, and HG6. Results from our comparative mapping study generally agreed with those of previous studies, i.e. the identification of clear macrosynteny over both entire genomes. However, our results suggest that partial rearrangement occurred more frequently between homeologous genomes in F. × ananassa and F. vesca than previously reported. In addition, genomic rearrangement across HGs was first observed in this study, i.e. between Fv-Chr2 and Fa-HG3 (Fig. 1). Of the LGs belonging to HG3, LG3A, 3B and 3C showed the genomic rearrangement, whereas it was not observed on LG3D.

The varieties of 129 F. × ananassa lines were distinguished to demonstrate the practicable utility of the markers developed in this study. Ninety-one percent (118 of 129) of the tested lines were developed in Japan. Because most Japanese lines were developed from a limited number of ancestral lines, such as ‘Haward17’ and ‘General Chanzy’, genetic diversity is considered to be generally narrow in Japanese germplasm.[49] Despite the expected narrow genetic diversity, the 45 SSR markers employed in this study identified most of the tested lines except for 2 pairs, ‘Nyoho’ and ‘Shin-Nyoho’ and ‘Toyonoka’ and ‘Himatsuri’. ‘Shin-Nyoho’ and ‘Himatsuri’ were developed from mutation lines of ‘Toyonika’ and ‘Nyoho’, respectively. Therefore, it was assumed that the genetic diversities were quite narrow in each pair of undistinguished varieties. Meanwhile, it is notable that the ‘Anter’, ‘Pihyaradondon’, and ‘Nyoho’ lines and the ‘Akita Berry’ and ‘Morioka 16’ lines could be distinguished from each other, although the former two and one of the varieties are mutant lines of ‘Nyoho’ and ‘Morioka 16’, respectively.

Variety identification is one of the major practical uses of DNA markers. Such identification should protect breeders' rights and inhibit the contamination of clonal seedlings during propagation. Because unknown samples are often used in variety identification, the accurate detection of targeted peaks is essential. In addition, it is well known that PCR results can vary depending on experimental conditions such as the type of thermal cycler and Taq polymerase employed during PCR analysis. In a previous study, Kunihisa[44] investigated the stability of polymorphic analysis of CAPS markers by comparing the results obtained in 14 laboratories to verify the adequacy of the makers used for variety identification tests in F. × ananassa. Govan et al.[50] also screened 32 SSRs that produced reliable PCR results for the identification of 60 varieties. In the present study, the 45 SSR markers were carefully screened by confirming the results of 3 independent PCRs. The 36 out of the 45 markers were mapped onto a total of 20 LGs of the integrated map. It suggested that the polymorphisms of selected markers roughly reflect genetic diversity in entire genome. Therefore, the markers were found to be reliable for variety identification.

In this study, we developed large numbers of SSR markers and constructed the densest linkage map for F. × ananassa. By performing comparative mapping and variety distinction, the resources that we developed were shown to be useful for genetic and genomic analysis as well as practical applications. Our results should contribute to the acceleration of advances in the study of F. × ananassa, along with the genus Fragaria.

Supplementary Data

Supplementary Data are available at www.dnaresearch.oxfordjournals.org.

Funding

This work was supported by the Kazusa DNA Research Institute Foundation, research and development projects for application in promoting new policies of agriculture, forestry and fisheries (21010) funded by the Ministry of Agriculture, Forestry and Fisheries.