Development and validation of allele-specific SNP/indel markers for eight yield-enhancing genes using whole-genome sequencing strategy to increase yield potential of rice, Oryza sativa L.

Background Rice is one of the major staple foods in the world, especially in the developing countries of Asia. Its consumption as a dietary source is also increasing in Africa. To meet the demand for rice to feed the increasing human population, increasing rice yield is essential. Improving the genetic yield potential of rice is one ideal solution. It is imperative to introduce the identified yield-enhancing gene(s) into modern rice cultivars for the rapid improvement of yield potential through marker-assisted breeding. Results We report the development of PCR-gel-based markers for eight yield-related functional genes (Gn1a, OsSPL14, SCM2, Ghd7, DEP1, SPIKE, GS5, and TGW6) to introduce yield-positive alleles from the donor lines. Six rice cultivars, including three each of donor and recipient lines, respectively, were sequenced by next-generation whole-genome sequencing to detect DNA polymorphisms between the genotypes. Additionally, PCR products containing functional nucleotide polymorphism (FNP) or putative FNPs for yield-related genes were sequenced. DNA polymorphisms discriminating yield-positive alleles and non-target alleles for each gene were selected through sequence analysis and the allele-specific PCR-gel-based markers were developed. The markers were validated with our intermediate breeding lines produced from crosses between the donors and 12 elite indica rice cultivars as recipients. Automated capillary electrophoresis was tested and fluorescence-labeled SNP genotyping markers (Fluidigm SNP genotyping platform) for Gn1a, OsSPL14, Ghd7, GS5, and GS3 genes were developed for high-throughput genotyping. Conclusions The SNP/indel markers linked to yield related genes functioned properly in our marker-assisted breeding program with identified high yield potential lines. These markers can be utilized in local favorite rice cultivars for yield enhancement. The marker designing strategy using both next generation sequencing and Sanger sequencing methods can be used for suitable marker development of other genes associated with useful agronomic traits. Electronic supplementary material The online version of this article (doi:10.1186/s12284-016-0084-7) contains supplementary material, which is available to authorized users.


Background
Food security is threatened by the growing human population and decreasing agricultural resources, including crop land and labor, through urbanization and industrialization. Current global yield increase rates (0.9-1.3% per year) of the four major crops (rice, maize, wheat, and soybean) are insufficient to meet food demand for the estimated nine billion people in 2050 (Khush 2005;Ray et al. 2013). Rice is a major staple food in the world, especially in the developing countries of Asia. In Africa, it is the most rapidly growing food source and, according to a conservative estimate, about 30 million tons more rice will be needed by 2035 (Seck et al. 2012). Within existing agricultural lands, the genetic improvement of yield potential in rice could be the ideal way to increase yield.
Three major traits (grain size, grain number per panicle, and panicle number per plant) are directly associated with rice grain productivity, and these traits strongly depend on the genetic potential of rice. However, these traits are complex and quantitative in nature. Through quantitative trait loci (QTL) analysis with fine mapping or positional cloning using rice mutants, about 20 genes that are involved in yield-related traits have been isolated in rice (Wang and Li 2005;Xing and Zhang 2010;Miura et al. 2011;Huang et al. 2013;Liu et al. 2015). In regards to the genes controlling grain size, six genes can be considered in rice breeding programs. The GW2 encoding RING-type E3 ubiquitin ligase regulates grain width. The loss of GW2 function by a premature stop codon caused by a 1-bp deletion on the fourth exon increased grain width, resulting in increased yield (Song et al. 2007). The GS3 gene encoding a putative transmembrane protein functions as a negative regulator for grain size. The nucleotide substitution from C to A on the second exon resulted in the stop codon (TGA) and eventually caused C-terminal truncated GS3 protein. This mutated allele increased grain size (Fan et al. 2006). The qSW5/GW5 gene encoding a nuclear-localized unknown protein is involved in the regulation of grain width. The loss-of-function allele by a 1.2-kb deletion including coding DNA sequence (CDS) of the gene increases grain width, and this allele is common in japonica varieties (Shomura et al. 2008;Weng et al. 2008). The GS5 gene encoding a putative serine carboxypeptidase regulates grain size. Different expression level of GS5 based on promoter sequences is associated with grain width . OsSPL16/qGW8, encoding SQUAMOSA promoter binding protein-like (SPL) 16, controls grain width, and higher expression in young panicle promotes grain width (Wang et al. 2012). The TGW6 gene encoding a novel protein having indole-3acetic acid (IAA)-glucose hydrolase activity regulates grain weight. The mutant allele (1-bp deletion in CDS) showed increased thousand-grain weight ).
Among the genes regulating grain number per panicle, several seem useful in a rice breeding program. The Gn1a gene encoding cytokinin oxidase/dehydrogenase2 (OsCKX2) regulates grain number per panicle. The nonfunctional allele of Gn1a increased grain number per panicle, resulting in increased yield (Ashikari et al. 2005). The Ghd7 encoding a CCT domain protein is involved in the regulation of heading date, plant height, and grain number per panicle. Fully functional alleles Ghd7-1 and Ghd7-3 showed delayed heading date and increased plant height and yield (Xue et al. 2008). The DEP1 gene encoding a phosphatidylethanolamine-binding proteinlike domain protein regulates panicle architecture. The mutant allele produced short panicle with high grain number per panicle, resulting in increased yield (Huang et al. 2009). The SPIKE/NAL1 gene encoding an unknown function protein regulates grain number per panicle. Either transcription level or three amino acids changes on NAL1 protein are associated with phenotype (Fujita et al. 2013). The SCM2/APO1 encoding F-box-containing protein controls grain number per panicle and culm diameter. Abundant transcripts of SCM2 in developing panicles increased grain number (Ookawa et al. 2010). Higher expression of OsSPL14/WFP/IPA1 gene encoding OsSPL14, in young panicle is associated with higher grain number per panicle (Jiao et al. 2010;Miura et al. 2010). Regarding the genes regulating panicle number per plant, several have been identified through positional cloning from the rice monoculm mutant moc1 and rice dwarf mutants with higher tiller number . However, the reported alleles of these genes may not be desirable in actual breeding programs because of too much reduction in tiller number of moc1 mutant or dwarfism in the high tiller number mutants. Some PCR markers for the yield-related genes were developed and used for marker-assisted breeding and gene function study. In the GS3 gene, the C/A single nucleotide polymorphism (SNP) on the second exon regarded as a functional nucleotide polymorphism (FNP) is directly associated with the phenotype, and it can be discriminated by digestion of PCR product with PstI restriction enzyme. Cleaved amplified polymorphic sequence (CAPS) markers for C/A SNP were reported (Fan et al. 2009;Yan et al. 2009;Wang et al. 2011). Additionally, Wang et al (2011) developed gene-tagged markers using the difference in repeat number of di-or tri-nucleotides within GS3 for marker-assisted selection (MAS). In the GW2 gene, the 1-bp deletion on the fourth exon is the FNP. This mutation was detected by the CAPS marker using HpaI restriction enzyme (Yan et al. 2009;Yan et al. 2011). For qSW5/GW5 gene, three alleles (Nipponbare type, Kasalath type, and indica II type) were discriminated by the N1212 marker, which used the difference in PCR product sizes among the alleles (Shomura et al. 2008;Yan et al. 2011). For the genes regulating grain number per panicle, the markers of Gn1a, OsSPL14, and DEP1 genes were reported. Yan et al (2009) developed the Gn1a-M1 marker for the detection of Gn1a-Habataki allele, which has a 16-bp deletion in the 5'-untranslated region (5'UTR), and the Gn1a-M2 marker for the detection of Gn1a-5150 allele, which has an 11-bp deletion on the third exon. Xu et al (2014) developed the CAPS marker using SduI restriction enzyme to select OsSPL14-Ri22 allele, which transcribes OsmiR156-resistant OsSPL14 transcripts caused by the nucleotide substitution at the OsmiR156 target site. They also developed the gene-tagged PCR marker of DEP1 gene using the 625-bp gap on the fifth exon between the yield-positive allele and non-target allele.
The allele-specific gene-tagged markers for the target genes are more effective than the genomic random markers surrounding the target gene (from several kb to a few Mb distance) because some markers will not show polymorphism in some recipient backgrounds and sometimes a false-positive allele can be selected by recombination between the target gene and the genomic random marker. However, allele-specific markers for yield-enhancing genes have not been developed yet for some major yield-enhancing genes or the existing markers need to be improved for an effective marker-assisted breeding program. The allele-specific gene-tagged markers were well developed for GS3, GW2, and qSW5/GW5 genes as mentioned above. But, for GS5, TGW6, SCM2, Ghd7, and SPIKE genes, PCR markers for selecting the yieldpositive alleles have not been developed yet. Additionally, the marker that can select the yield-positive OsSPL14-ST12 allele, which is different from the OsSPL14-Ri22 allele, has not been reported. For the Gn1a gene, a 16-bp deletion in 5'UTR is common in indica rice varieties (54.3%) (Yan et al. 2009), suggesting that the Gn1a-M1 marker may not function in many indica varieties. For the DEP1 marker , the intensity of the nontarget allele band was very weak compared with that of the target allele in the heterozygous plants. To obtain clear genotype data, the previous DEP1 marker needs to be improved.
Although around 20 yield-related genes have been identified in rice, their evaluation and use are still limited for rice breeding programs. These studies are very important for improving the genetic yield potential of rice effectively. To evaluate and use the yield enhancing genes, in this study we report allele-specific gene-tagged markers for eight yield-related genes. For effective screening of DNA polymorphisms of yield-enhancing genes, we employed both next-generation sequencing as well as Sanger sequencing methods. For easy access and user-friendly markers, we developed PCR-gel-based markers without restriction enzyme digestion. In addition, for high-throughput genotyping using a fluorescence detection system, we also developed Fluidigm SNP genotyping markers. Moreover, the allele-specific PCR-gel based markers were validated through its application in our marker-assisted selection (MAS) breeding program.

Results and Discussion
Development of allele-specific markers for eight yieldenhancing genes was conducted as a part of our breeding program for increasing genetic yield potential in 12 elite indica varietal backgrounds (Table 1). For determination of alleles of yield-enhancing genes in 12 recipient cultivars and effective screening of polymorphisms between the donor and the recipient, we sequenced the whole genomes of three donors (Habataki, ST6, and ST12) and three recipients (NSIC Rc158, NSIC Rc222, and NSIC Rc238) using next-generation sequencing technique. In addition, the target yield-enhancing genes were also sequenced from the 12 recipients and the donors using Sanger sequencing method.
Generally, several DNA polymorphisms such as indel and/or SNP were found between the two parental alleles as a result of fine mapping of major QTLs. Among polymorphisms, the FNP was identified in some yield-enhancing genes such as GS3, GW2, DEP1, and TGW6 genes but not in some other genes. For marker designing, firstly we targeted FNP or putative FNPs of yield-related genes. Secondly, we selected the donor-specific polymorphisms within the target gene to select the donor allele easily. Actually, when breeders have the original donor source which was used for QTL/fine-mapping analysis, any DNA polymorphisms within or near the target gene between the donor and the recipient can be used to design a marker for MAS (Usually indel polymorphisms are preferred to design markers because of convenience to marker designing and clear genotype results by the markers). The selected polymorphisms for each yieldrelated gene and its corresponding markers were described below one by one. Finally, the markers we developed were validated through its application in our MAS breeding lines.

Gn1a Markers
Yield-positive Gn1a allele was originally isolated from the high-yielding rice variety Habataki through QTL analysis with positional cloning (Ashikari et al. 2005) and Gn1a-Habataki allele was also found in cultivar ST12 ). These two materials were used as a Gn1a donor in our marker-assisted breeding program. We screened DNA variations in the Gn1a gene, including in the 5-kb upstream region from the start codon and 2-kb downstream region from the stop codon, using IGV software (Robinson et al. 2011) with WGS data. Within this range, the sequences were the same among Habataki, ST6, ST12, and NSIC Rc222, indicating that these four accessions have the yieldpositive Gn1a allele. Three SNPs in the promoter region were found based on WGS analysis (Additional file 1: Figure S1A) and the sequences of three SNPs were confirmed by Sanger sequencing of PCR products from Gn1a donors and 12 recipients (Fig. 1a). Based on the sequence analysis, all genotypes were divided into three groups (Types 1-3; Fig. 1a). NSIC Rc222, IR05N412, CT5805, and Irga427 among the 12 recipients have Gn1a-Habataki allele (Type 3). First, we developed the Gn1a-indel1 marker to detect a 16-bp indel in 5'UTR as identified by Ashikari et al (2005). Except for Parao, all genotypes showed the same band size as Habataki ( Fig. 1b bottom), indicating that this marker is available only for Gn1a MAS in a Parao background. Next, we developed the Gn1a-17 SNP marker using the tetraprimer PCR method (Additional file 2: Figure S2A) to discriminate G/A SNP in the promoter region (Fig. 1a). The marker discriminated G/A SNP clearly in the parents (Fig. 1b top) as well as in intermediate breeding lines that contained heterozygous plants (Fig. 1c). This marker is available to introduce Gn1a-Habataki allele (Type 3) into both Gn1a-Type 1 and Gn1a-Type 2 backgrounds. Additionally, we developed an indel-type marker, the Gn1a-indel3 marker, based on WGS data analysis, which showed approximately 70-bp indel polymorphism near 3'UTR (Additional file 1: Figure S1B). This marker was tested in the parental varieties, and the result was consistent with the Gn1a-17 SNP marker analysis, except for Parao ( Fig. 1b middle), indicating that the marker is available to introduce Gn1a-Habataki allele into Gn1a-Type 2 backgrounds but not Parao background. The Gn1a-indel3 marker was also applied to intermediate breeding lines and the genotyping results were consistent with those of the Gn1a-17 SNP marker (Fig. 1c). Based on the Gn1a sequence context of the background varieties, a suitable marker among the Gn1a-indel1, Gn1a-indel3, and Gn1a-17 SNP markers can be used to select the yield-positive Gn1a allele. Reduced expression of Gn1a in inflorescence meristem of Habataki elevates cytokinin content, resulting in high grain number per panicle. Therefore, we tried to find polymorphisms in the Gn1a promoter region, which is the regulatory region for gene transcription. Through WGS and PCR product sequencing of the Gn1a gene, we found three SNPs in the promoter region. It was presumed that any of the three SNPs were associated with Gn1a expression level. Based on the combination of the three SNPs and 16-bp indel in the Gn1a promoter region, we assigned rice accessions into three Gn1a alleles (Types 1-3) (Fig. 1a). Earlier the effect of the Gn1a-Habataki allele was evaluated in only japonica backgrounds (Type 1) (Ashikari et al. 2005;Ohsumi et al. 2011). When 35 indica varieties were analyzed by the Gn1a-M1 marker that discriminates the 16-bp indel in the 5'UTR, 19 accessions (54.3%) showed a Habataki band pattern (Yan et al. 2009), suggesting that these lines are Type 2 or Type 3 alleles. In our analysis, 11 out of 12 recipients showed 16-bp deletion in the 5'UTR like Habataki. This suggested that many indica varieties belong to Type 2 or Type 3 alleles, unlike japonica varieties. However, the effect of the Gn1a-Habataki allele (Type 3) needs to be tested in Type 2 backgrounds through the development of near-isogenic lines (NILs). If the Type 3 allele is functionally the same as the Type 2 allele (i.e., no grain number difference between the recipient variety and NIL-Type 3), rice breeders do not need to introduce the Gn1a-Habataki allele into Type 2 cultivars. The previous Gn1a-M1 marker is not available to introduce the Habataki allele in Type 2 backgrounds because it cannot discriminate the Type 2 allele and Type 3 allele. The Gn1a-17SNP marker and three Fluidigm SNP markers (Gn1a-17SNP-FD, Gn1a-18SNP-FD, and Gn1a-19SNP-FD) are effective for introducing the Gn1a-Habataki allele into Gn1a-Type 1 background as well as Type 2 background. Regarding detection of the 16-bp indel, we newly designed the Gn1a-indel1 marker, which enhanced the band separation in agarose gel by reducing PCR product sizes (99 bp and 115 bp) compared with those of the Gn1a-M1 marker (113 bp and 129 bp). The Gn1a-indel3 marker located near the 3'UTR was developed for obtaining clear genotype data. The marker showed clear band separation in agarose gel and similar PCR efficiency between the two alleles in heterozygotes (Fig. 1c). Although the Gn1a-indel3 marker was not associated with the phenotype, it will be helpful for MAS of the Gn1a gene in many indica cultivar backgrounds.

OsSPL14/WFP/IPA1 Markers
The QTL WFP (WEALTHY FARMER'S PANICLE) encoding OsSPL14 negatively regulates the number of tillers in the vegetative stage and positively controls the number of rachis in the reproductive stage. Higher expression of OsSPL14 in young panicles increased panicle branching and grain number per panicle, resulting in increased yield. Two different yield-positive alleles (OsSPL14-ST12 and OsSPL14-Aikawa1) showed higher expression in inflorescence with different mechanisms. In the line ST12, OsSPL14 transcripts were abundant by less DNA methylation in the OsSPL14 promoter region compared with that of Nipponbare rather than by DNA sequence differences ). The OsSPL14-ST12 allele belongs to epigenetic alleles that show heritable gene expression difference which is caused not by DNA sequence variations but by DNA methylation or chromatin status (Kakutani 2002). In contrast, the The sequence alignment showed nucleotide polymorphisms in the promoter and 5'UTR regions of Gn1a among 15 varieties, including the reference genome (Nipponbare), two Gn1a donors, and 12 recipients. DNA polymorphisms were highlighted with pink and green color. The number of unrepresented nucleotides (bp) in a sequence was shown in parentheses. Based on the context of the promoter sequences, three Gn1a alleles (Types 1-3: T1-T3) were found. Variety name with asterisk (*) was used as the donor line of target allele. b Agarose gel images analyzed by three Gn1a markers (Gn1a-17SNP, Gn1a-indel3, and Gn1a-indel1 markers) from two donor lines and 12 recipients. Predicted PCR product sizes for yield-positive allele (P), non-target allele (N), and common band (OP) were shown at the right side of the gel image. Primer combination for each marker was shown on the gel images and its sequences were listed in Table 2. M, DNA size marker. c Application of Gn1a markers in the intermediate breeding line. Fourteen BC 1 F 3 plants derived from PR37951 x Habataki cross were genotyped with Gn1a-17SNP and Gn1a-indel3 marker, respectively. Genotyping result was scored for each plant as PP (homozygous for positive allele), NN (homozygous for non-target allele), and PN (heterozygote) OsSPL14-Aikawa1 allele expressed OsmiR156-resistant OsSPL14 transcripts because of the nucleotide substitution (C to A) at the OsmiR156 target site located on the third exon of OsSPL14, resulting in higher expression in panicles . The same alleles with Aikawa1 were isolated through fine mapping of QTL IPA1 (Ideal Plant Architecture 1) from the japonica lines Shaoniejing and Ri22 (Jiao et al. 2010). Through WGS data analysis, we found a putative ST12-specific SNP (C/T) in the OsSPL14 promoter region (Additional file 3: Figure S3). Finally, this was confirmed by direct sequencing of PCR product from line ST12 and 12 recipients (Fig. 2a). The SNP was selected for developing the SPL14-04SNP marker which consists of two PCRs: the C-allele primer set (SPL14-04SNP-F/CR) and the T-allele primer set (SPL14-04SNP-F/TR). The SPL14-04SNP marker showed allele-specific PCR amplification in parental lines (Fig. 2b) and in intermediate breeding lines harboring heterozygous plants (Fig. 2c). The Aikawa1-specific SNP located at OsmiR156 target site was confirmed by Sanger sequencing from Aikawa1 and 12 recipients (Fig. 2a). The SNP was used for developing the SPL14-12 SNP marker to select OsSPL14-Aikawa1 allele. The non-Aiakawa1 allele and Aikawa1 allele were amplified with SPL14-12SNP-CF/R primer set and SPL14-12SNP-F/AR primer set, respectively. The marker discriminated the alleles clearly from the parents (Fig. 2b) and intermediate breeding lines (Fig. 2d).
The OsSPL14-ST12 allele is an epigenetic allele. Less DNA methylation in OsSPL14 promoter in ST12 increased its transcription, resulting in high grain number per panicle. However, we can introduce the OsSPL14-ST12 allele using DNA polymorphism within or near the OsSPL14 gene between ST12 and the recipient instead of checking the status of DNA methylation. Among the six varieties analyzed by WGS, Habataki had many unique DNA polymorphisms (18 SNPs and one 2-bp deletion) in the OsSPL14 gene located on chromosome 8 (data not shown). But, the new OsSPL14 allele, Habataki allele, may not be a new yield-positive allele because the QTL controlling grain number per panicle was not detected on chromosome 8 (Ashikari et al. 2005). However, we found an ST12-specific C/T SNP in the promoter region, which was about 4.2 kb distance from the start codon. This nucleotide was identical to the border nucleotide of the 2.6kb WFP candidate region ), indicating Fig. 2 OsSPL14 markers. a Identification of donor-specific DNA polymorphisms in OsSPL14 gene. The donor-specific SNPs at the promoter and the third exon were represented through DNA sequence alignment among 15 varieties, including Nipponbare, two different OsSPL14 donors (ST12 and Aikawa1), and 12 recipients. The location of 2.6-kb candidate region of WFP ) was depicted. b Agarose gel images analyzed by the SPL14-04SNP marker and the SPL14-12SNP marker, respectively, from the parental lines. Primer combination for each PCR was shown on the gel images. c Application of SPL14-04SNP marker in the intermediate breeding line, with 12 BC 2 F 2 plants derived from PR38012 x ST12 cross. d Application of SPL14-12SNP marker in the intermediate breeding line. Twelve BC 2 F 2 plants derived from IR04A115 x Aikawa1 cross were genotyped that the SNP can be an FNP of the OsSPL14-ST12 allele. In plants, DNA methylation occurs through the addition of a methyl group to the cytosine bases of DNA in the context of CG, CHG, and CHH (H = A, C, or T) ). The nucleotide substitution from C to T resulted in CCC (other varieties) and TCC (ST12) sequence context, respectively (Fig. 2a). If this C base is a target site of methylation, the T base of ST12 can avoid methylation, indicating higher OsSPL14 expression. Miura et al (2010) tested DNA methylation levels in the 2.6-kb WFP candidate region using bisulfite sequencing and the result showed some differences in the region between ST12 and Nipponbare, which is about 1070-bp distance from the C/ T SNP locus. Further study is needed to show that this SNP has some function.
The marker for the OsSPL14-ST12 allele has not been reported. We developed the SPL14-04SNP marker using the ST12-specific SNP to select the OsSPL14-ST12 allele. This marker will be available for most indica and japonica background varieties because the SNP is quite unique in ST12. To select OsSPL14-Aikawa1/Shaoniejing/Ri22 allele, Xu et al. (2014) developed the CAPS marker using SduI restriction enzyme. Here, we introduced the allele-specific PCR-gel-based marker, the SPL14-12SNP marker. This marker will be effective for MAS to most of rice varieties because the OsSPL14-Aikawa1 allele is rare.

SCM2/APO1 Marker
ABERRANT PANICLE ORGANIZATION1 (APO1) encodes an F-box-containing protein and is involved in controlling rachis branching in panicle, tiller outgrowth, and culm diameter. The apo1 mutant formed smaller inflorescences with reduced numbers of branches and spikelets (Ikeda et al. 2007). In contrast, the APO1-overexpressing mutant, Undulate rachi1 (Ur1) and APO1overexpressing transgenic plants showed dramatic increment of grain number per panicle but tiller number per plant reduced remarkably (Murai and Iizawa 1994;Ikeda-Kawakatsu et al. 2009). STRONG CULM2 (SCM2) is a mild allele of APO1 found in variety Habataki (2-3 times higher expression of APO1 in inflorescence than that of Koshihikari) and the allele increased culm diameter and grain number per panicle without a reduction in tiller number (Ookawa et al. 2010;Terao et al. 2010). This SCM2-Habataki allele is regarded as a useful allele of APO1 gene for increasing yield and lodging resistance in a breeding program. Through WGS data analysis, we tried to find Habataki-specific DNA variations in the SCM2 gene (-5 kb from start codon and +2 kb from stop codon) among the six cultivars. However, no sequence difference was found among the five genotypes (Habataki, ST12, NSIC Rc158, NSIC Rc222, and NSIC Rc238). To determine the APO1 allele from 12 recipients, we sequenced two promoter regions and the third exon (Fig. 3a), which showed some DNA variations between genotypes of mapping bi-parents (Habataki and Koshihikari) (Ookawa et al. 2010). Except for Parao, 11 recipients had the same sequence as Habataki. However, the 12-bp indel in the promoter region (Fig. 3a) was selected for marker development to introduce the SCM2-Habataki allele into Parao. We developed the SCM2-indel1 marker and tested it in the parental lines. As expected, only Parao showed a different band size (Fig. 3b). The marker was also applied in intermediate breeding lines that were BC 3 F 2 plants derived from a Parao x Habataki cross and it distinguished the genotypes clearly (Fig. 3c). For automated highthroughput electrophoresis, the same PCR product was analyzed using Fragment Analyzer (Fig. 3d).
The WGS and PCR product sequencing showed that the SCM2-Habataki allele was detected in most of our donors and recipients. These results indicate that the high-yielding and lodging-resistant SCM2-Habataki allele is widely dispersed in indica varieties. In contrast, temperate japonica varieties Nipponbare, Koshihikari, Sasanishiki, and Aikawa1 had the non-yield-positive SCM2 allele, suggesting that the SCM2-Habataki allele will be effective in many temperate japonica varieties. The SCM2-indel1 marker may be helpful for MAS in japonica backgrounds. The SCM2 sequence of ST6 was almost the same as that of Nipponbare but two 2-bp deletions in the promoter region (3.4-kb and 5.8-kb regions from the start codon) were found in only the line ST6 among the six lines in the WGS analysis carried out. The new allele, SCM2-ST6, needs to be tested if the ST6 allele is superior to the Habataki allele through checking SCM2 expression level in stem, young panicle, and phenotypes in NIL-ST6 because the grain number per panicle and culm diameter of ST6 were higher than those of Habataki as observed at IRRI during the 2013-2015 experiments.

Ghd7 Marker
Grain number, plant height, and heading date7 (Ghd7) encoding a CCT domain protein is involved in multiple yield-related traits, including heading date, plant height, and grain number per panicle (Xue et al. 2008). Nine Ghd7 alleles (Ghd7-0~7 and Ghd7-0a) were identified based on the predicted protein sequences from about 120 genotypes (Xue et al. 2008;Lu et al. 2012). To know the allele type of Ghd7 from our breeding materials, we analyzed the genomic sequence (protein coding regions and an intron) of Ghd7 using WGS data of six genotypes. ST6 variety had Ghd7-6 allele and the others had exactly the same DNA sequences as variety Minghui63, which had the fully functional yield-positive Ghd7-1 allele. Additionally, we sequenced PCR products containing Ghd7 coding sequences from Habataki, ST12, ST6, and 12 recipients, resulting in that all lines (except for ST6 and Parao) had the Ghd7-1 allele. These results indicated that 11 out of 12 recipients already have the yield-positive Ghd7-1 allele. However, we selected the A/T SNP (521st nucleotide from ATG in CDS) and developed the Ghd7-05SNP marker detecting the SNP to introduce the Ghd7-1 allele into Parao which had the Ghd7-2 allele (a functional but weaker allele). Genotyping results from parental lines through this marker were consistent with DNA sequencing data, indicating that the marker discriminated A (non-target allele) and T (Ghd7-1 allele) clearly (Fig. 4a). To replace the Ghd7-2 allele with the Ghd7-1 allele in the Parao background, this marker was used in intermediate breeding lines derived from a Parao x Habataki cross (Fig. 4b).
Based on our sequencing data analyses of Ghd7, the Ghd7-1 allele is common in indica rice varieties. This result was consistent with the previous reports (Xue et al. 2008;Lu et al. 2012) that Ghd7-1 and Ghd7-2 alleles were popular in indica and japonica varieties, respectively. So, the fully functional Ghd7-1 allele may be effective for increasing yield potential in many japonica varieties which are growing in long-day condition. The Ghd7-05SNP marker will be helpful to introduce Ghd7-1 allele into japonica backgrounds. Thirteen BC 3 F 2 plants derived from Parao x Habataki cross were genotyped by the SCM2-indel1 marker and its PCR products were electrophoresed in 4 % agarose gel (c) and in Fragment Analyzer (d) Fig. 4 Ghd7 marker. a Agarose gel image analyzed by Ghd7-05SNP marker from the parental lines. Non-target allele and yield-positive allele (Ghd7-1) were amplified by Ghd7-05SNP-F/AR primer pairs and Ghd7-05SNP-F/TR primer pairs, respectively. Habataki and ST12 were used as the donors of Ghd7-1 allele and Aikawa1 (japonica-type variety) was included as a control of Ghd7-2 allele. b Application of Ghd7-05SNP marker in the intermediate breeding line. Thirteen BC 2 F 2 plants crossed between Parao and Habataki were genotyped by the marker DEP1 Marker DENSE AND ERECT PANICLE1 (DEP1) encodes a phosphatidylethanolamine-binding protein-like domain protein and regulates panicle architecture. The natural DEP1 mutant allele (625-bp deletion on the fifth exon) encoding C-terminal truncated DEP1 protein increased grain number per panicle, resulting in increased yield, although panicle length decreased. This allele might be derived from Italian landrace Balilla and was distributed to many high-yielding Chinese japonica varieties (Huang et al. 2009). In order to detect FNP, the forward primer and the reverse primer of the DEP1-indel1 marker were located upstream and downstream of the 625-bp deletion on the fifth exon, respectively (Fig. 5a). As a result, the difference in PCR product sizes will be 625 bp between the yield-positive allele and the non-target allele. The Turkish high-yielding variety Osmancik-97 was used as the donor of DEP1. The DEP1-indel1 marker was tested in the donor and 12 recipients, and it discriminated DEP1 alleles clearly (Fig. 5b). But, when it was applied to intermediate breeding lines, the band intensity of the non-target allele (1031 bp) was markedly lower than that of the positive allele (406 bp) in heterozygotes (plant #2, 3, 5, 6, 7, 10, and 12 in Fig. 5d). To solve the problem of the marker, we developed the tri-primer PCR method, which contained the two primers of DEP1-indel1 marker (DEP1-indel1-F/R) and additional nontarget allele-specific primer (DEP1-indel1-625F) (Fig. 5a). As a consequence, the band size of the non-target allele decreased from 1031 bp to 310 bp. The new marker (DEP1-indel1P) worked well in parental lines (Fig. 5c) as well as in heterozygotic plants from intermediate breeding lines (Fig. 5e).
Like our DEP1-indel1 marker, the previously reported DEP1 marker ) also showed similar results: that the marker was no problem in homozygous plants but the PCR band intensity was quite different between two alleles in heterozygotes. This phenomenon was caused by the difference in PCR amplification efficiency between two different PCR products in the same reaction tube during PCR cycling. Usually, this occurs when the gap between two PCR products is large (several hundred base pairs) such as the DEP1-indel1 marker (625-bp gap) or the indel region contains difficult PCR sequences such as high G-C sequence or inverted repeat sequence. Generally, the longer PCR product and the difficult PCR region show lower amplification efficiency, resulting in a weaker band in agarose gel. To achieve similar PCR efficiency between nontarget allele and yield-positive allele, we reduced the gap from 625 bp to 96 bp by adding DEP1-indel1-625F primer (Fig. 5a), resulting in a tri-primer PCR system (DEP1-indel1-F/R/625F). Eventually, the DEP1-indel1P marker generated similar band intensity between nontarget allele and positive-allele bands in heterozygotes (Fig. 5e). The marker will be more effective for MAS of the DEP1 gene than previous marker.

SPIKE/LSCHL4/NAL1/GPS Markers
SPIKELET NUMBER (SPIKE), LSCHL4, and GPS are allelic to Narrow leaf1 (NAL1) encoding a plant-specific protein with unknown biochemical function (Qi et al. 2008;Fujita et al. 2013;Takai et al. 2013;Zhang et al. 2014). The NAL1 gene was originally identified through characterization of a classic rice dwarf mutant, nal1 (Qi et al. 2008), and the other alleles were characterized by QTLs with map-based cloning by independent research groups. The SPIKE allele from tropical japonica landrace Daringan and the LSCHL4 allele from temperate japonica variety Nipponbare increased grain number per panicle and grain yield in indica backgrounds (Fujita et al. 2013;Zhang et al. 2014). The protein coding sequences of NAL1 were identical among japonica varieties (Daringan, Nipponbare, and Koshikari) and among indica varieties (IR64, 93-11, and Takanari), respectively. The NAL1japonica allele and the NAL1-indica allele were distinguished by three SNPs causing amino acid change in CDS (Fujita et al. 2013;Takai et al. 2013;Zhang et al. 2014). Based on the sequence analysis of NAL1 gene, 16 cultivars were classified into five haplotypes (Takai Our WGS data showed that Habataki, NSIC Rc222, and NSIC Rc238 had the same DNA sequences as the NAL1-Takanari allele (Type 1) (Additional file 4: Figure S4). In the case of NSIC Rc158, several SNPs were newly found but the deduced amino acids were the same as those of the NAL1-Takanari allele (Type 1). So, these three recipients may need the NAL1-japonica allele to increase yield potential. We developed three allelespecific PCR markers: the SPIKE-01SNP marker detecting G/A SNP on the third exon, the SPIKE-03SNP marker detecting G/A SNP on the fifth exon, and the SPIKE-indel3 marker detecting a 20-bp indel in the promoter region (1.9 kb distance from the start codon). These three markers were tested in the donor line and 12 recipients (Fig. 6a-c). Among the 12 recipients, CT5805 and Parao showed the same band pattern as the SPIKE donor line (YP9), suggesting that the two varieties had the yield positive NAL1-japonica allele (Type 5). Therefore, we need to introduce the NAL1-Type 5 allele into 10 recipients. The breeding line (BC 2 F 1 ) derived from a PSB Rc82 x YP9 cross was analyzed by the three SPIKE markers (Fig. 6d-f). As expected, the three markers showed the same results among markers. In conclusion, the SPIKE markers are effective for introducing a Type 5 allele in an indica background.
The yield-positive NAL1 allele originated from japonicatype varieties and this allele increased yield in several indica-type cultivars (Fujita et al. 2013;Zhang et al. 2014). In the 12 recurrent parents, only two lines (CT5805 and Parao) had the yield-positive NAL1 allele. This result indicates that the NAL1-japonica allele will be effective in many indica cultivars for increasing yield. Here, we developed three gene-tagged markers, SPIKE-01SNP marker, SPIKE-03SNP marker, and SPIKE-indel3 marker, discriminating the japonica allele and indica allele. These markers will be useful for MAS of NAL1 alleles. Through WGS data analysis, we found new two haplotypes (Type 6 and Type 7) to the NAL1 gene (Additional file 4: Figure S4). Line ST6 producing high grain number per panicle needs to be tested to determine whether it is superior to or has the same allele as the japonica allele because an additional amino acid change was found on the fifth exon in ST6.

GS5 Markers
Abundant transcripts of GS5, encoding a putative serine carboxypeptidase, in the hull (palea/lemma) before heading and in the developing endosperm enhanced grain width and grain weight. Through sequence analysis of the GS5 promoter (2 kb upstream from the start codon) from 35 cultivars, the sequences were divided into three groups: wide grain (WG) allele, medium grain (MG) allele, and narrow grain (NG) allele . First, we checked the GS5 alleles using WGS data analysis. The sequences of GS5 of Habataki, ST12, NSIC Rc222, and NSIC Rc238 were the same as those of GS5-NG allele. NSIC Rc158 and ST6 belonged to the GS5-MG allele and GS5-WG allele, respectively. Through PCR product sequencing of GS5 promoter region, GS5 alleles were determined from 12 recipients. Parao belonged to the GS5-WG allele and NSIC Rc158, PR38012, CT5803, CT5805, and Irga427 belonged to the GS5-MG allele. Based on the sequence alignment of GS5 promoter, we selected the MG-specific 4-bp indel and the WGspecific C/T SNP to design the GS5-indel1 marker and the GS5-03SNP marker, respectively (Fig. 7a). The GS5-indel1 marker produced a 63-bp band (MG allele) and a 67-bp band (WG or NG). The 4-bp difference could not be separated properly in agarose gel. After PCR amplification from the parental lines, the PCR products were analyzed by high-resolution electrophoresis such as PAGE (data not shown) or capillary electrophoresis using Fragment Analyzer (Fig. 7b). The genotyping Fig. 6 SPIKE markers. a-c Agarose gel images analyzed by three SPIKE markers from the parental lines. The SPIKE-01SNP marker discriminated G/A SNP located on the third exon using two separated PCRs (SPIKE-01SNP-GF/R primer set and SPIKE-01SNP-AF/R primer set) (a). The SPIKE-03SNP marker detected the G/A SNP on the fifth exon using two separated PCRs (SPIKE-03SNP-F/GR primer set and SPIKE-03SNP-F/AR primer set) (b). The SPIKE-indel3 marker distinguished a 20-bp indel located in the promoter region (c). YP9 was used as a donor line of SPIKE gene. d-f Application of the SPIKE-01SNP marker (d), the SPIKE-03SNP marker (e), and the SPIKE-indel3 marker (f) in the intermediate breeding line. Thirteen BC 2 F 1 plants derived from PSB Rc82 x YP9 cross were genotyped by the three markers, respectively result was consistent with the DNA sequencing data. To select the WG allele, we developed the GS5-03SNP marker. Genotyping PCR was performed with four primers (GS5-03SNP-OPF/OPR/TF/CR) from the parental lines and electrophoresis was done in agarose gel (Fig. 7c). The genotype result was consistent with sequencing data, indicating that the marker data were reliable. The marker was also validated in intermediate breeding line and discriminated SNPs properly (Fig. 7d).
The GS5 promoter sequences of our parental lines were classified into the existing three GS5 alleles (WG, MG, and NG) ). In the 12 recurrent parent lines, the numbers of WG, MG, and NG alleles were one, five, and six accessions, respectively. This result indicated that many indica varieties had MG or NG alleles. To increase grain size, WG or MG alleles of GS5 need to be introduced in indica cultivars. The GS5-indel1 marker and the GS5-03SNP marker will be helpful for this purpose. For the GS5-03SNP marker, the largest band (499 bp) generated by common outer primers was weaker than the allele-specific bands (257 bp and 291 bp) (Fig. 7c-d). This could be caused by the difference in PCR efficiency between/among bands in one tube. However, this is not a problem because there was no band intensity difference between two allele-specific bands in heterozygotes (Fig. 7d).

TGW6 Marker
TGW6, encoding a novel protein having hydrolase activity from IAA-glucose into IAA and glucose, regulates grain weight through controlling both source ability and sink size. TGW6 is a single-exon gene and the loss-offunction allele caused by a 1-bp deletion in CDS in Kasalath increased grain weight . We developed the TGW6-1d marker detecting FNP (1bp deletion) to select a yield-positive allele (TGW6-Kasalath allele). The marker consisted of two sets of primers (TGW6-1d-F/NR for the non-target allele and TGW6-1d-F/PR for the yield-positive allele) and it was tested in the donor line (Kasalath) and 12 recipients. All genotypes were determined clearly by this marker and only Kasalath had the yield-positive TGW6 allele among our breeding materials (Fig. 8a). The marker also functioned properly in intermediate breeding lines (Fig. 8b).
The loss-of-function allele of the TGW6 gene caused by a 1-bp deletion enhanced grain weight. Among the 69 cultivars, only four varieties had the TGW6-Kasalath allele, suggesting that the yield-positive allele is a rare allele . A similar result was observed in our study. All 12 recipients had the non-yield-positive TGW6 allele. Therefore, the TGW6-Kasalath allele may be effective for increasing grain weight in most cultivars throughout indica-and japonica-type varieties. The TGW6-1d marker detecting the FNP will be useful for introducing the TGW6-Kasalath allele through MAS.

High-throughput Genotyping Markers Using Fluidigm SNP Genotyping Platform
Currently, several high-throughput genotyping platforms using a fluorescence detection system are available in rice (Thomson 2014). We also developed fluorescence-labeled allele-specific markers using Fluidigm SNP genotyping platform for high-throughput MAS of yield-related genes. For the Gn1a gene, three SNPs in the promoter region (Fig. 1a) were discriminated by Gn1a-19SNP-FD, Gn1a-18SNP-FD, and Gn1a-17SNP-FD markers (Fig. 9). For the OsSPL14 gene, the ST12-specific SNP in the promoter region and the Aikawa1-specific SNP on the third exon ( Fig. 2a) were detected by SPL14-04SNP-FD marker and SPL14-12SNP-FD marker, respectively. In the Ghd7 gene, the 521st nucleotide (A/T) in CDS was used for developing the Ghd7-05SNP-FD marker. We developed the GS5-01SNP-FD marker distinguishing T (NG allele)/C (MG or WG allele) SNP in the GS5 promoter region (1594-nucleotide upstream from the start codon of H94 line, Gen-Bank accession no. JN256057). And, for selection of the GS5-WG allele, the GS5-03SNP-FD marker was developed. For the GS3 gene, FNP (C/A SNP) on the second exon was used for developing the GS3-01SNP-FD marker. All markers were tested in 47 samples consisting of 39 accessions ( Fig.9; Additional file 5: Table S1). Among these, ST12 and Aikawa1 were unique at the SPL14-04SNP-FD marker and SPL14-12SNP-FD marker, respectively (Fig. 9). The genotyping results were consistent with the DNA sequencing data as well as the PCR-gel-based allele-specific markers developed in this study.
In marker-assisted breeding, breeders sometimes handle several hundred samples and/or many markers. Usually, MAS should be finished before flowering to make a backcross or a cross between different gene sources for gene pyramiding. In this case, high-throughput genotyping systems are very helpful. Here, we employed Fluidigm SNP genotyping system. In this system, 24-192 samples can be analyzed with 24-96 customized markers simultaneously. After PCR, fluorescence from the allelespecific PCR products can be detected without gel electrophoresis, and data calling can be computed rapidly. We developed nine Fluidigm SNP markers for five yieldrelated genes. These markers will be useful for highthroughput genotyping of yield-related genes in rice.

Conclusion
Here, we described the procedures of SNP/indel marker development for yield-enhancing genes, including allele determination of target genes in the recurrent parental lines, DNA polymorphism screening using WGS and Sanger sequencing techniques, the selection of DNA polymorphisms for marker designing, and marker validation. This procedure can be applied to designing markers for other agronomic traits. For rapid improvement of genetic yield potential in rice, we need to evaluate and use the previously identified yield-enhancing genes in local favorite cultivar backgrounds. The allelespecific markers for yield-enhancing genes shown in this study will be very helpful for MAS of yield-enhancing traits/genes. Through whole-genome sequencing, we identified some new haplotypes for some yield-enhancing genes such as the SCM2-ST6 allele and SPIKE-ST6 allele. For the same gene, the alleles superior to the previously identified alleles can be found. Therefore, if the new allele was found in an advanced phenotype line such as ST6, the effect of the new allele needs to be tested. To increase rice productivity, the identification of new yield-enhancing genes, isolation of superior alleles of previously identified genes, marker development of the genes for MAS, evaluation of the effect of the genes/alleles, and their use in actual breeding programs should be performed rapidly and updated continuously.

Plant Materials
Habataki, ST12, ST6, Aikawa1, YP9, Osmancik-97, and Kasalath varieties were used as donors of yield-positive alleles of yield-related genes. In addition, 12 elite indica varieties originating from the International Rice Research Institute (IRRI), Philippines, and International Center for Tropical Agriculture (CIAT), Colombia, were used as recipients to increase genetic yield potential through the introduction or pyramiding of yield-enhancing genes ( Table 1). Seeds of the donor lines Habataki, ST12, ST6, Aikawa1, and Kasalath were obtained from the Bioscience and Biotechnology Center, Nagoya University, Japan, and seeds of Osmancik-97 were obtained from Trakya Agricultural Research Institute, Edirne, Turkey. The plants were grown in the IRRI paddy field, Laguna, Philippines, in 2012-2015. To introduce the yield-positive alleles of yield-related genes, crosses were made between the donor lines and recipients followed by selfing, backcrossing, and crossing with other cross combinations in the same background, based on the breeding plan. The markers we developed were applied to intermediate breeding lines for MAS.

Whole-genome Sequencing and Comparative Sequence Analysis
In order to identify nucleotide polymorphisms of yieldenhancing genes between the donors and recipients, WGS was performed using Illumina HiSeq 2000 by the Yale Genomics Center (Connecticut, USA) for three donors (Habataki, ST6, and ST12) and three recipients (NSIC Rc158, NSIC Rc222, and NSIC Rc238). The sequencing results yielded 609,301,492 reads in all. Total mapped reads ranged from 47.65 % to 57.69 % and from 50.29 % to 58.42 % against Nipponbare and 93-11 genomes, respectively. Uniquely mapped reads were on average 42.01 % (Additional file 6: Table S2). The short reads were aligned against the Nipponbare genome (MSU v7). The SNP and indel polymorphisms were screened using IGV software (Robinson et al. 2011; https://www.broadinstitute.org/igv/). Fig. 9 Genotyping of five yield-related genes from 39 rice accessions using nine Fluidigm SNP genotyping markers. The marker name with the detected-nucleotide polymorphism was shown on the top of each graph. The allele-specific primers were labeled with HEX fluorescence dye (green) for the yield-positive allele and with FAM fluorescence dye (red) for the non-target allele. After PCR on Fluidigm array, the alleles of samples were determined by fluorescent signal intensity. In each marker result, the yield-positive allele of the genes was shown with a green letter. Black spot, non-template control; gray spot, invalid call; blue spot, heterozygote

PCR Product Sequencing
To confirm DNA polymorphisms obtained by WGS and to determine the alleles of yield-related genes from 12 recipients, the PCR products containing putative FNPs or donor-specific polymorphisms for yield-related genes were sequenced by the Sanger method. Genomic DNAs were extracted from 12 recipients and four donors (Habataki, ST12, ST6, and Aikawa1) using the CTAB method ) and PCR amplification was performed for Gn1a, OsSPL14, SCM2, Ghd7, and GS5 genes. The primers for preparation of PCR product and DNA sequencing are listed in Additional file 7: Table S3. DNA sequencing reactions were performed with BigDye terminator sequencing kit v3.1 (Applied Biosystems, www.appliedbiosystems.com/), and then the sequences were determined by Applied Biosystems 3730xl DNA analyzer by Macrogen Inc., Republic of Korea (www.macrogen.com/).

Development of Allele-specific Markers
The DNA sequences of yield-related genes were obtained from NCBI GenBank (www.ncbi.nlm.nih.gov/) using the accession codes presented in the original papers (Ashikari et al. 2005;Ikeda et al. 2007;Xue et al. 2008;Huang et al. 2009;Miura et al. 2010;Li et al. 2011;Ishimaru et al. 2013;Takai et al. 2013). To find DNA polymorphsims between the donors and recipients, multiple sequence alignments were performed using BioEdit software (Hall 1999). Special DNA variations such as FNP, putative FNP, donor-specific polymorphisms, and polymorphisms between the donors and recipients were selected for marker designing. For indel variation, the forward and reverse primers covered the indel region. When the gap was small (less than 20 bp), we reduced PCR product size (<120 bp) to increase the degree of band separation in agarose gel. To discriminate SNPtype DNA polymorphisms, first we used the tetra-primer PCR method (Ye et al. 2001). This type of marker can reveal a SNP by one PCR reaction with four primers. The Gn1a-17 SNP marker and GS5-03 SNP marker were developed based on this method. Secondly the separated allele-specific PCR method was used to design the other markers. This method requires two separate PCR reactions with allele-1-specific primer and allele-2-specific primer, respectively, to define the SNP. Actually, we tried the tetra-primer PCR method for all SNP polymorphisms because of its convenience, but the results from some testing markers were unstable among trials. So, we accepted the separated allele-specific PCR method. To increase allele-specific primer annealing, an artificialmismatched nucleotide was given near the 3' end of the allele-specific primer (Hayashi et al. 2004). For better understanding, the designing of the Gn1a-17 SNP marker and the SPIKE-01 SNP marker are shown in Additional file 2: Figure S2. The primer information for allele-specific markers is presented in Table 2.

DNA Preparation, PCR with Allele-specific Markers, and Electrophoresis
Leaf samples were collected from the intermediate breeding lines in the paddy field. Fresh or stored leaf (-20°C) was cut at about 4 cm long and was placed into a 2-mL tube with a steel ball. Then, the tubes were frozen in liquid nitrogen and leaf tissues were ground with the 2010 Geno/Grinder (www.spexsampleprep.com). In each tube, 200 μL of DNA extraction buffer (100 mM Tris-HCl, pH 9.5, 1 M KCl, 10 mM EDTA, pH 8.0) was added and the samples were incubated at 65°C for 30 min. The samples were diluted by adding 1 mL of water and centrifuged for 10 min at the maximum speed. The supernatant was transferred into a 96-well plate and stored at 4°C for further uses. All genotyping markers followed the PCR conditions mentioned below. The 20-μL PCR solution contained 1× PCR buffer, 200 μM of each dNTP, 0.25 μM of each primer, 1.5 μL of supernatant (template DNA), and 1 unit of Taq DNA polymerase. Thermal cycles were programmed as follows: 94°C, 3 min; 35 cycles of 95°C for 25 s; 55°C for 25 s; and 72°C for 35 s, concluding with 72°C for 5 min. The PCR products amplified with Gn1a-indel1 marker, SCM2-indel1 marker, and SPIKE-indel3 marker were electrophoresed in 4 % agarose gel and the others in 2.5 % agarose gel. In the case of GS5-indel1 marker, its PCR products (63 bp/67 bp) could not be separated properly in agarose gel. So, it was analyzed by 8 % polyacrylamide gel electrophoresis (PAGE) or capillary electrophoresis.

High-throughput Capillary Electrophoresis
Gel electrophoresis requires some labor and time because of manual gel preparation and hand loading of PCR products. Currently, automated high-throughput capillary electrophoresis instruments are available and are helpful for handling a huge number of samples. Here, we used Fragment Analyzer (Advanced Analytical Technologies Inc., www.aati-us.com/) that supports ready-to-use polymer (gel), automated sample loading, and data calling. The PCR products produced by the GS5-indel1 marker and the SCM2-indel1 marker were analyzed by Fragment Analyzer.

Fluidigm SNP Genotyping
We selected nine SNPs from five yield-related genes to design Fluidigm SNP genotyping markers. The sequences, including both 250-bp upstream and 250-bp downstream of the target SNP, were sent to Fluidigm (Assay_Design_Group@fluidigm.com) and the Fluidigm SNP genotyping markers consisting of specific target amplification (STA) primer, locus-specific (LS) primer,