ORIGINAL RESEARCH published: 07 August 2020 doi: 10.3389/fpls.2020.01205 Resynthesis: Marker-Based Partial Reconstruction of Elite Genotypes in Clonally-Reproducing Plant Species Iban Eduardo 1,2, Simó Alegre 1,3, Konstantinos G. Alexiou 1,2 and Pere Arús 1,2* 1 IRTA (Institut de Recerca i Tecnologia Agroalimentàries), Barcelona, Spain, 2 Centre for Research in Agricultural Genomics (CRAG), CSIC-IRTA-UAB-UB, Campus UAB, Barcelona, Spain, 3 Parc Cientı´fic i Tecnològic Agroalimentari de Lleida, Lleida, Spain We propose a method for marker-based selection of cultivars of clonally-reproducing plant species which keeps the basic genetic architecture of a top-performing cultivar (usually a partly heterozygous genotype), with the addition of some agronomically relevant differences (such as production time, product appearance or quality), providing added value to the product or cultivation process. The method is based on selecting a) two complementary nearly-inbred lines from successive selfing generations (ideally only F2 and F3) of large size, that may generate individuals with most of their genome identical to the Edited by: original cultivar but being homozygous for either of the two component haplotypes in the Inaki Hormaza, rest, and b) individuals with such characteristics already occurring in the F2. Option a) Institute of Subtropical and allows for introgressing genes from other individuals in one or both of these nearly-inbred Mediterranean Horticulture La Mayora, Spain lines. Peach, a woody-perennial, clonally-reproduced species, was chosen as a model for Reviewed by: a proof of concept of the Resynthesis process due to its biological characteristics: self- Amy Iezzoni, compatibility, compact and genetically well-known genome, low recombination rates and Michigan State University, United States relatively short intergeneration time (3–4 years). From 416 F2 seedlings from cultivar Sweet Eiji Yamamoto, Dream (SD), we obtained seven individuals with 76–94% identity with SD, and selected Kazusa DNA Research Institute, Japan five pairs of complementary lines with average homozygosity of the two parents ≥0.70 *Correspondence: such that crossing would produce some individuals highly similar to SD. The application of Pere Arús pere.arus@irta.cat this scheme to other species with more complex genomes or biological features, including its generalization to F1 hybrids, is discussed. Specialty section: This article was submitted to Keywords: Breeding methods, Complementary lines, nearly-inbred lines, gene introgression, vegetatively- Plant Breeding, reproducing crops, peach, F1 hybrids a section of the journal Frontiers in Plant Science Received: 24 May 2020 INTRODUCTION Accepted: 24 July 2020 Published: 07 August 2020 A few high-quality cultivars dominate the market of many clonally propagated species that are often Citation: highly, or at least partly, heterozygous individuals. Sometimes they are recognized by their name Eduardo I, Alegre S, Alexiou KG and and their main features by the consumers (for example cultivars in grape, apple and almond). Arús P (2020) Resynthesis: Marker- Reproducing the genotype of a top performing cultivar with a minor but significant improvement Based Partial Reconstruction of Elite Genotypes in Clonally-Reproducing may have a high market value (e.g., a change in bloom or maturity time, skin or flesh fruit color, Plant Species. flesh sugar or acidity content and composition, and postharvest behavior). With classical Front. Plant Sci. 11:1205. approaches (sexual crossing and progeny selection) this is extremely unlikely, and to our doi: 10.3389/fpls.2020.01205 knowledge, it has never been tried. Other approaches have difficult-to-overcome regulatory issues Frontiers in Plant Science | www.frontiersin.org 1 August 2020 | Volume 11 | Article 1205 Eduardo et al. Resynthesis and sometimes a negative social perception (genetically modified MATERIALS AND METHODS or gene-edited materials), or are rare and occur unpredictably (somatic mutations). Resynthesis is an approach based on Breeding Scheme conventional breeding technologies that, with the help of Resynthesis A molecular markers, allows the selection of two homozygous (or This method consists of selecting two highly homozygous and nearly homozygous) complementary individuals that, when complementary individuals in the selfed offspring of the original crossed, produce offspring with a similar, albeit different, variety that, when crossed, reproduce most of its heterozygous genetic composition from the original cultivar. Homozygous make-up along with additional homozygous regions (Figure 1A). fragments of either of the two parents in specific genome regions that were heterozygous in the original cultivar could result in phenotypic changes for traits of interest in the context of A the performance of a well-defined variety. In addition, the development of two parental lines allows for the incorporation of an exotic gene (disease resistance or fruit quality trait) in one of them and the Resynthesis of the improved cultivar in a predictable time frame (two generations after the first cross). The number of crossovers per chromosome and meiosis is limited, usually between one and three in plants (Mercier et al., 2015), and the frequency of non-recombined chromosomes is high enough to find certain specific combinations (fully homozygous or heterozygous chromosomes) in a selfed offspring of reasonable size (Koumproglou et al., 2002). The basis of Resynthesis is that, with the help of markers covering the whole genome, it is possible to select fully homozygous and complementary genotypes to reproduce (resynthesize) an initial partly heterozygous genotype (i.e., a cultivar) with a limited number of selfing or backcross generations. The “Reverse Breeding” method (Dirks et al., 2009) is a similar concept that was successfully tested in Arabidopsis (Wijnker et al., 2014). However, this approach involves the use of genetically modified lines with suppressed recombination and the production of doubled-haploid lines by in vitro culture, two methods that are well established in a limited number of crops. Additionally, Reverse Breeding was proposed to generate homozygous parents to obtain a hybrid identical to one already existing, B while our goal is the generation of individuals with a similar genetic background but with different characteristics with respect to a top performing clonally-reproducing cultivar. This strategy is particularly suitable for clonally reproducing species, fruit trees, berries, and ornamentals, among others. Key elements of their biology are the perennial habit of most species and a long intergeneration period, often accompanied by a juvenile period, which requires as few generations as possible to obtain a new cultivar. Our model, peach, has certain advantages: a relatively short intergeneration period (3–4 years), self- compatibility, and a compact, diploid and sequenced genome (250 Mbp; x=8) (Arú s et al., 2012; Verde et al., 2013), with a low recombination rate (1.2–1.5 crossovers/chromosome and meiosis) and a high level of homozygosity in the commercial cultivars (Micheletti et al., 2015; da Silva Linge et al., 2018). In this paper we FIGURE 1 | Breeding schemes for a species of eight pairs of chromosomes of: develop the basic scheme for Resynthesis, and its genetic (A) Resynthesis A, where two nearly homozygous and complementary lines are foundation, and provide experimental data from a trial using selected to obtain an individual highly similar to the original cultivar; (B) Resynthesis the peach “Sweet Dream” (SD) cultivar. Due to the length of time B, generation of a collection of individuals nearly identical to a peach cultivar, but required for proof of concept, here we provide only data from the with a small number of chromosomal segments in homozygosis. All individuals of the collection presented have five heterozygous chromosomes as in the original marker selection of SD selfed progeny to evaluate the progress cultivar and only three homozygous fragments each. made and the feasibility of the approach. Frontiers in Plant Science | www.frontiersin.org 2 August 2020 | Volume 11 | Article 1205 Eduardo et al. Resynthesis This approach has the advantage that the two selected parents, genome) are monomorphic (Donoso et al., 2015; Serra et al., being highly homozygous, can be used for the introgression of 2017), identity by descent (IBD) being the most logical cause due new characters from other sources. The disadvantage is that to a high level of co-ancestry (Micheletti et al., 2015) of peach obtaining these two complementary lines may typically require commercial cultivars. This determines that some of the crossovers more than one generation of selfing plus an additional generation produced at the homozygous DNA fragments will not be effective. to obtain the resynthesized line (or set of lines) and evaluate As an approximation, if we consider that an individual with 50% its behavior. IBD has a continuous fragment of half of the length of each chromosome in homozygosis, the probability of a fully Resynthesis B homozygous chromosome with a single recombination event Using the same F2 population as for Resynthesis A, it is possible to increases to 9/32, and the probabilities of having eight or five or select plants with a similar composition to the original cultivar more chromosome pairs fully homozygous can be estimated as (Figure 1B). For this, markers are used to select individuals with 3.9×10−5 and 4.5×10−2, respectively (Table 1). For a finer most chromosomes the same as the original variety and others with estimation including recombination rates, we know that in peach more homozygous regions, so that the new homozygous regions the number of crossovers per chromosome (co/cs) per meiosis are a low proportion of the genome (5–20% of the heterozygous ranges between 1 and 2, and an overall estimation of 1.2–1.5 is genome in homozygosity). The objective is to evaluate the behavior reasonable (Arú s et al., 2005; da Silva Linge et al., 2018). We have of the selected plants in the field, and their value as potential new calculated approximate probabilities of finding a given number of varieties. Additionally, such individuals may constitute an chromosomes in homozygosis (or heterozygosis) considering 1.2 interesting resource for genetic studies, consisting of plants and 1.5 co/cs: 1.2 (80% of the meiosis with a single co/cs and 20% containing different homozygous regions in the otherwise partly with two) and 1.5 (50% with one co/cs and 50% with two), and heterozygous background of the original variety. 50% IBD (half a chromosome in homozygosis and half in heterozygosis). These calculations indicate that the expected Genetic Basis of the Resynthesis Process probabilities are compatible with finding some of the desired The proportions of different genotypes expected in an F2 progeny genotypes in populations of several hundred plants (Table 1). were calculated for peach with x=8 chromosomes, initially Considering as an example a peach F2 progeny of N=2,000, assuming one crossover per chromosome. In this case, the with 1.2 co/cs and 50% homozygosity, an average of nine plants probability of having one given chromosome pair in homozygosis with six or more homozygous chromosome pairs would be is 1/8, so the probability of the eight chromosome pairs being expected, or 50 plants with five or more homozygous homozygous is (1/8)8≈6.0×10−8, or 1.2×10−3 for five or more chromosomes. In this case, the objective would be to select 1–5 homozygous chromosome pairs (Table 1; Figure 1) calculated pairs of plants with four or more chromosomes in homozygosis from a binomial distribution with p=1/8. The probabilities are the and complementary to each other. This is achievable because the same for finding individuals entirely heterozygous like the original number of individuals needed to find two complementary cultivar, for one or more chromosomes. lines from a set of lines with four homozygous chromosomes In the case of peach, there is an additional advantage: recent is n=10 with a 0.95 probability. This is calculated with the results have shown that vast regions of the chromosomes of the formula n(n−1)/2=log0.05/log[(2 x −1)/2 x ], where n is the modern cultivars (sometimes accounting for 40-60% of the number of individuals with x (four in this case) homozygous chromosomes. And approximately one in four of the possible TABLE 1 | Approximate probabilities of finding a progeny homozygous (or pairs of these 50 plants with five homozygous chromosomes heterozygous) for various intact chromosomes (cs) in the selfed progeny of peach would have four homozygous chromosomes in common. It is (x=8). also important to be sure that the three non-homozygous Peach chromosomes of each pair of lines allow construction of two complementary lines without homozygous fragments of the Fully heterozygous 50% of the genome identical by same haplotype in both lines. By selfing these individuals, sets parent; 1 co/cs1 descent2 of genotypes could be obtained that are homozygous for six or 1 co/cs 1.2 co/cs 1.5 co/cs more chromosomes and complementary in a sample of N=1,000. From these we would expect to obtain at least one pair of lines All cs in 6.0×10−8 3.9×10−5 1.3x10−5 2.1x10−6 homozygosis that, crossed, would produce a plant ≥ 95% genetically identical All cs in 6.0×10−8 3.9×10−5 1.3x10−5 2.1x10−6 to SD with >10% probability. heterozygosis We could use the same F2 population of N=2,000 to test the Six or more 8.5×10−5 8.0×10−3 3.8×10−3 1.1×10−3 Resynthesis B approach by selecting for heterozygous individuals homozygous cs with chromosomal compositions identical to SD, with the same Five or more 1.2×10−3 4.5×10−2 2.5×10−2 9.4×10−3 homozygous cs probabilities as those calculated for finding a homozygous Four or more 1.1×10−2 0.16 0.11 5.2×10−2 genotype (Table 1). A set of 50 plants on average with five or homozygous cs more chromosomes identical to SD can be selected. This set 1 co/cs, crossovers/chromosome; 2Estimated assuming that half of each chromosome is would contain plants having a few homozygous fragments in an heterozygous and the other half homozygous. essentially heterozygous background allowing genetic analysis of Frontiers in Plant Science | www.frontiersin.org 3 August 2020 | Volume 11 | Article 1205 Eduardo et al. Resynthesis the segregation in these regions by comparison with the original be heterozygous in SD and previously validated in other peach cultivar. Plants of this set can be selected to have the ensemble of materials from the peach 9k array (Verde et al., 2012; Micheletti homozygous fragments covering ideally the whole genome for et al., 2015). For genomic regions not covered by these markers, each of the two haplotypes, a situation somewhat similar to a near- we identified nine more heterozygous SNPs in SD from the isogenic line (NIL) collection in a heterozygous background. Some resequencing data. The complete SNP list with their main of these lines could be advanced to new cultivars. characteristics is given in Tables S1 and S2. Fresh leaves from SD and from its F2 progeny were collected Case Study: Resynthesis of Cultivar directly in 96 well plates. DNA was extracted following the Doyle “Sweet Dream” (SD) and Doyle (1990) protocol. Genotyping was carried out with the Plant Materials KASPar SNP Genotyping System (KBiosciences, Herts, UK) Fruit from a single stand commercial plantation of SD were adapted to a Biomark™ HD platform (Fluidigm) and using the collected at maturity in a commercial orchard of Els Alamú s universal KASPar MasterMix (LGC, England) following the (Lleida) in July of 2015, 2016 and 2017 (4,300 fruit in total). instructions recommended by the supplier. KASPar assay Seeds were extracted, cleaned and kept in vacuum bags until primers were designed with Primer Picker (KBiosciences) November, when they were stratified at 4˚ C for three months, (Table S1), and SNP data were analysed with the Fluidigm after which they were germinated in trays in the greenhouse. SNP Genotyping Analysis software (Fluidigm). Germinated plants were transplanted to the field of the IRTA Markers with less than 90% of the data available were Experimental Station of Gimenells, and grown following discarded. An initial marker order was established according to standard agricultural practices. their physical position and SNPs located on the same chromosome were placed in the same phase. Missing data were Resequencing “Sweet Dream” imputed in chromosome fragments that were delimited by Young leaves were collected in the spring of 2015 and DNA markers with the same genotype. Imputations were limited to extracted for resequencing using DNeasy® Plant Mini Kit 1.6% of the data available. Markers were then mapped with (Qiagen) following the manufacturer’s instructions. A paired- Kosambi’s genetic distance using MapMaker 3.0 (Lander et al., end sequencing library with an average insert size of 315 bp was 1987), and MapChart 2.1 (Voorrips, 2002) was used to draw the prepared following the manufacturer’s instructions (Illumina). resulting linkage map. The library was sequenced at CNAG (Centre Nacional d’Anàlisi Only mapped markers were used for the selection for Genòmica, Barcelona) using the Illumina HiSeq2500 platform at Resynthesis A and B. For each plant, we calculated the a paired-end mode with 2x250 cycles. Adapter removal and read frequency of heterozygous loci and ordered the plants according trimming was done using Trimmomatic v0.36 (Bolger et al., to this parameter. For selection of plants with complementary 2014). FastQC (version 0.10.1; https://www.bioinformatics. genotypes, we identified first the plants with heterozygosity values babraham.ac.uk/projects/fastqc/) was used for quality control below 0.25 and looked for possible complementary pairs by of the reads before and after adapter removal and trimming. selecting those with higher distance values among those with With a good quality dataset, reads were mapped onto the peach heterozygosity values below 0.50. Distance was calculated for each genome downloaded from GDR (ftp://ftp.bioinfo.wsu.edu/ locus as 0 for markers with the same genotype, 0.5 for species/Prunus_persica/Prunus_persica-genome.v2.0.a1/ heterozygous vs. homozygous comparisons and 1 for the two assembly/Prunus_persica_v2.0.a1_scaffolds.fasta.gz) using alternative homozygotes and summed over all loci. Only pairs of BWA-MEM (v0.7.5a) (Li and Durbin, 2009). Alignment files genotypes with high distances that were totally compatible, were sorted and reads that were uniquely mapped to the genome i.e. where the same allele was not fixed at any given locus for were retained for variant calling using Samtools v1.3.0 (Li, 2011), both parents, were accepted as possible parents for the taking into account reads with minimum base quality of 10 and next generation. minimum mapping quality of 20. From the list of variants, we removed INDELs, sites where the quality of the SNP and the genotype quality were lower than 30, sites with depth lower than RESULTS 12, sites with more than one alternative allele and, finally, sites where there was a ratio imbalance >0.3 between forward and From the approximately 4,300 seeds stratified, 416 (135 in 2015, reverse reads of the same allele. In order to keep sites with high 136 in 2016 and 145 in 2017), germinated and produced viable genotype qualities we removed those sites where the heterozygous plants. This is a low germination rate (~10%), considering that SNPs had an allele frequency (AF) ratio outside the 0.4 – 0.6 range we used the standard methods of IRTA’s peach breeding and the homozygous SNPs had an AF ratio lower than 0.8. program, where the average germination rate is 20–30%, although it is known that certain parents and crosses produce SNP Array Development, Genotyping, very poor yields, and that there is an extremely wide range of Mapping and Selection in the SD F2 success (0–90%), depending on the cross. One of the reasons we Population selected SD was because, being a mid-season producing cultivar, To develop a 96 SNP array to genotype the SD F2 seedlings, SNPs no need for “in vitro” embryo rescue was expected, as occurs in covering the whole genome were selected. We chose 87 known to early varieties. We attribute this low rate of germination to the Frontiers in Plant Science | www.frontiersin.org 4 August 2020 | Volume 11 | Article 1205 Eduardo et al. Resynthesis characteristics of the cultivar, requiring specific adaptation of the eight Prunus chromosomes. The order of the markers within the germination protocol to improve results. linkage groups was identical to their physical order in the Illumina sequencing of SD DNA generated 91,441,354 reads chromosomes (see Table S1). The map covered a total distance with a total output of 22.8Gbp. After quality filtering we were left of 422.5 cM, ranging from 36.0–63.7 cM/chromosome and the with 81,866,684. Of these, 66,950,399 (81.8%) were mapped onto markers studied covered 71% of the total physical length of the the peach genome, covering 94% of its length with at least five peach genome. Two of the main regions uncovered by the map, reads. The resulting average depth was determined to be 36X. In end of chromosome 1 (21.3 Mbp) and end of chromosome 7 (4.2 total, we detected 94,451 heterozygous SNPs, with chromosomes Mbp), were predominantly homozygous (Figure S1) in SD. Pp02 having the highest and Pp08 the lowest number, and with For Resynthesis A, we selected four pairs of complementary lines an average of 11,806 heterozygous SNPs per chromosome (Table with average heterozygosity of the two parents ≤0.29 (Figure 2A). S3). This corresponds to an average SNP density of 418.7 SNPs/ These lines had on average 3.3 chromosomes in homozygosity Mbp, with Pp02 and Pp01 having the highest (726.1) and lowest ranging from 0 to 8, and the average number of estimated (144.5) density, respectively. Looking at the distribution of recombinations per genome was 4.8 (0.6 co/cs), ranging from 2 to heterozygous SNPs along the chromosomes, they are 8. Recovering entire chromosomes without any recombinations concentrated in certain parts of the genome, leaving large would be the ideal situation, because the presence of one or more interspersed homozygous regions, particularly in chromosomes recombination events almost necessarily implies that there will be a 1, 3, 6, and 8 (Figure S1). Considering that regions below 5% of more or less large homozygous region in the resynthesized genotype, the average of SNPs/Mbp (21 SNP/Mbp; see Figure S1) are as crossovers rarely occur at the same genomic site in two meiosis. potentially identical by descent, we calculate that the genome of For Resynthesis B, we found seven individuals with ≥73% SD would be 23% IBD, in agreement with the high level of heterozygosity (range 73–94%) that contained the full genome of inbreeding of peach cultivars (Micheletti et al., 2015). SD and a few DNA fragments that were fixed (Figure 2B). One From the 96 SNPs used for analysis, 64 segregated and were of them (SD039) contained a single homozygous fragment, two mapped and of the remaining 32, 16 were monomorphic and 16 contained three and the remaining four plants, four. Two of the had no call or an unclear segregation pattern. Of the 64 mapped lines, SD208 and SD075, were homozygous in the regions of SNPs, 62 were codominant with the three expected genotype chromosomes 2 and 4, respectively, both regions known to classes and two (SNP_IGA_671806 and SNP_IGA_682254) were contain gene clusters affecting various agronomic characters in dominant, with only two genotypes. A map was constructed that Prunus (Aranzana et al., 2019), which are likely to produce detected all eight linkage groups corresponding to the expected phenotypes of special interest. A B C FIGURE 2 | Chromosome diagram of “Sweet Dream” F2 selected individuals. G1-G8 peach linkage groups. He: proportion of heterozygous loci; MHe: mean He for each pair of complementary lines. # Int: introgression number (A) Pairs of nearly-homozygous complementary lines selected for Resynthesis A; with high probability their offspring will produce individuals very close to ‘Sweet Dream’. (B) Selected individuals for Resynthesis B, nearly identical to ‘Sweet Dream’. (C) Fully homozygous F2 plants. Frontiers in Plant Science | www.frontiersin.org 5 August 2020 | Volume 11 | Article 1205 Eduardo et al. Resynthesis When looking for individuals with high levels of homozygosity, seed multiplication before grafting techniques became sufficiently we found two that were completely homozygous for all markers well established. Growing a number of progeny from self- studied (Figure 2C). One of them (SDF2-238) was selected as one compatible cultivars of other fruit trees (almond, cherry, apricot) of the parents of the complementary lines. Both plants had two could produce inbreds that may be of use for genetic analysis, recombination events, meaning that in the regions where including ‘de novo’ sequencing. recombination occurred there would be a small heterozygous Massive linkage disequilibrium is a favorable circumstance DNA fragment, the rest of the genome being fully homozygous. for Resynthesis and species with low chromosome numbers and low recombination rates, usually associated with small genomes (Mercier et al., 2015), are good candidates for a fast Resynthesis process. To measure haplotype integrity after meiosis, we DISCUSSION propose a recombination index Ri = n(1+CpC), where n is the number of chromosomes and CpC the average number of A major use of molecular marker analysis is to predict the crossovers per chromosome that can be inferred from linkage phenotype based on genotypic data, which has important maps available. Species with low Ri would require less individuals applications in plant breeding. One of the critical advantages and/or less generations to obtain suitable pairs of inbreds for of this strategy is that, once the marker-phenotype association Resynthesis. Peach, for example, with eight chromosomes and has been established, it is possible to genotype large progenies at 1.3 co/cs (da Silva Linge et al., 2018) would have a Ri =8+8x1.3 = the seedling stage, and only a few selected individuals need to be 18.4. Almond (x=8, 1.0 co/cs) Ri=16.0 (Sá nchez-Pé rez et al., grown and phenotyped. In the case of Resynthesis, the target 2007), raspberry (x=7, 1.3 co/cs) Ri=16.1 (Ward et al., 2013) or genotype is known and the objective is to obtain, in the shortest citrus (x=9, 1.7 co/cs) Ri= 24.3 (Huang et al., 2018) are examples possible timeframe, individuals having a similar genetic makeup, of favorable organisms for Resynthesis, whereas apple (x=17 1.5 some maintaining most of its good phenotypic qualities, but with co/cs) Ri=42.5 (Di Pierro et al., 2016) or grape (x=19; 1.5 co/cs) commercially relevant differences. These differences would be Ri= 47.5 (Tello et al., 2019), would require greater population initially provided by the inter-allelic variability of the same sizes to obtain sufficiently close individuals. CpC for each species cultivar, but in addition, Resynthesis A would give access to has been estimated as the total distance of its map given in the alleles of a broader germplasm pool by introgressing them in the reference cited divided by 50 cM/crossover multiplied by the inbred parents that would again generate the target genotype. haploid chromosome number. The capability of each species to In spite of the relatively low number of seedlings analyzed, we produce large numbers of seedlings per generation at affordable found four pairs of complementary lines which were suitable for costs is also important: in this respect grapes and apple are more a second generation of Resynthesis A. We expect to perform the amenable to Resynthesis than peach. Hybrid populations first crosses with these lines in the spring of 2021 to obtain between pairs of complementary lines could be obtained using populations that will be genotyped and integrated in our partly homozygous lines from early steps of selfing (i.e. from F2 breeding program for the following years, to select for selected couples), and plants with valuable phenotypes selected at individuals similar to SD, with specific characters incorporated. this stage. Selfing generations to fix specific lines could be done in These four pairs of parents will also be selfed, allowing for parallel. Backcrossing of selected lines to obtain parents with additionally fixed (5–6 chromosomes in homozygosity) and genes from other materials (disease resistance or others), could complementary individuals to be selected. also be started at early stages. We found seven individuals nearly identical to SD It is important that the genotype to be resynthesized is self- (Resynthesis B), which gives us an opportunity to look for the compatible as the complementary lines are obtained from effects of specific regions of the genome on key agronomic successive selfing generations. For self-incompatible genotypes, characters, although to a limited extent (six genomic regions in a possible way to circumvent the problem is crossing first with a total covering chromosome 6 and parts of chromosomes 1, 5, 7, self-compatible genotype, select the self-compatible individual in and 8). More individuals will be produced with appropriate this progeny with the least number of recombination events in compositions from the crosses between the complementary the original cultivar gamete, and start the selection from the BC1 pairs of Resynthesis A that, once genotyped, will allow for between the hybrid line and the target genotype. Then only self- additional genome coverage. It would not be surprising if some compatible lines having as much as possible of their genome of these individuals have characteristics that may already make fixed for one of the alleles of the target genotype (looking for them commercially interesting for selection as new cultivars. pairs of complementary lines following a model similar to The two completely homozygous lines were an unexpected Resynthesis A) or that are highly similar to its genetic makeup result, showing that these individuals can be recovered with a (Resynthesis B), should be selected. limited number of offspring. One of them was selected as one of Another aspect to consider is the existence of lethal alleles in the parents for Resynthesis A, but beyond that, these lines are of heterozygosis in the original cultivar. This cannot be discarded in biological interest as a relatively rare event. The self-pollinating highly heterozygous genomes of typically outcrossing species, nature of peach provides the opportunity to produce inbreds, such as those of almond and grape, especially in old cultivars of and certain traditional cultivars of Spain are highly homozygous unknown origin. In this case, one of the two complementary (Li et al., 2013; Micheletti et al., 2015) as a consequence of their lines must be kept in heterozygosis for the smallest possible Frontiers in Plant Science | www.frontiersin.org 6 August 2020 | Volume 11 | Article 1205 Eduardo et al. Resynthesis genomic region around the gene causing the lethal effect, AUTHOR CONTRIBUTIONS and selection should be for heterozygous individuals in the final cross between the two nearly-homozygous lines in the last IE performed the proof of concept experiment, obtained and step of Resynthesis A. Knowing whether a certain cultivar has analyzed the SNP data and revised the manuscript. SA developed lethal or deleterious alleles in heterozygosis could be tested the population for the experiment and revised the manuscript. by constructing a linkage map with markers in the F2 and KGA, analyzed the DNA resequence data and revised the identifying genomic regions with highly distorted segregation manuscript. PA designed the method, supervised the experiment frequencies. No regions of this type have been detected in our F2 and drafted and revised the manuscript. population of SD. Resynthesis could be expanded to other crops, particularly those that are seed propagated and where F1 hybrids are the FUNDING usual cultivar type. In this case, there is an alternative approach to the usual generation of inbred lines and selection of the best This research was supported in part by grants from: The Spanish performing F1 hybrids produced between them: whenever the Ministry of Economy and Competitiveness (MINECO/FEDER evaluation of single genotypes (individually or as a set of clonally projects AGL2012-40228-C02-01 and RTA2015-00050-00-00), propagated plants) from segregating progenies can be done with the project RIS3CAT (COTPA-FRUIT3CAT) financed by the sufficient accuracy, highly performing individuals could be European Regional Development Fund through the FEDER selected and their genotype A resynthesized, following a frame of Catalonia 2014-2020, the Severo Ochoa Program for strategy that somewhat mimics the clonally-propagated Centres of Excellence in R&D 201-2019 (SEV-2015-0533) and breeding scheme. CERCA Programme-Generalitat de Catalunya. This manuscript has been released as a pre-print at BioRxiv (Eduardo et al., 2020). DATA AVAILABILITY STATEMENT SUPPLEMENTARY MATERIAL Novel resequencing data are deposited in ENA (European Nucleotide Archive) with sample ID: SAMEA5704508. SNPs The Supplementary Material for this article can be found online for the genotyping are all described in detail in Supplementary at: https://www.frontiersin.org/articles/10.3389/fpls.2020.01205/ Table S2. full#supplementary-material REFERENCES Eduardo, I., Alegre, S., Alexiou, K. G., and Arú s, P. (2020). Resynthesis: Marker- based partial reconstruction of elite genotypes in clonally-reproducing plant Aranzana, M. J., Decroocq, V., Dirlewanger, E., Eduardo, I., Gao, Z. S., Gasic, K., species. bioRxiv. 1–24. doi: 10.1101/2020.05.09.085738 et al. (2019). Prunus genetics and applications after de novo genome Huang, M., Roose, M. L., Yu, Q., Du, D., Yu, Y., Zhang, Y., et al. (2018). sequencing: achievements and prospects. Hortic. Res. 6, 58. doi: 10.1038/ Construction of High-Density Genetic Maps and Detection of QTLs s41438-019-0140-8 Associated With Huanglongbing Tolerance in Citrus. Front. Plant Sci. 9, Arú s, P., Yamamoto, T., Dirlewanger, E., and Abbott, A. G. (2005). “Synteny in the 694. doi: 10.3389/fpls.2018.01694 Rosaceae,” in Plant Breeding Reviews, vol. 27. Ed. J. Janick (Hoboken, NJ: John Koumproglou, R., Wilkes, T. M., Townson, P., Wang, X. Y., Beynon, J., Pooni, H. S., Willey & Sons, Inc.), 175–211. et al. (2002). STAIRS: a new genetic resource for functional genomic studies of Arú s, P., Verde, I., Sosinski, B., Zhebentyayeva, T., and Abbott, A. G. (2012). The Arabidopsis. Plant J. 31 (3), 355–364. doi: 10.1046/j.1365-313X.2002.01353.x peach genome. Tree Genet. Genomes 8, 531–547. doi: 10.1007/s11295-012- Lander, E. S., Green, P., Abrahamson, J., Barlow, A., Daly, M. J., Lincoln, S. E., et al. 0493-8 (1987). MAPMAKER: An interactive computer package for constructing Bolger, A. M., Lohse, M., and Usadel, B. (2014). Trimmomatic: a flexible trimmer primary genetic linkage maps of experimental and natural populations. for Illumina sequence data. Bioinformatics 30, 2114–2120. doi: 10.1093/ Genomics 1, 174–181. doi: 10.1016/0888-7543(87)90010-3 bioinformatics/btu170 Li, H., and Durbin, R. (2009). Fast and accurate short read alignment with da Silva Linge, C., Antanaviciute, L., Abdelghafar, A., Arú s, P., Bassi, D., Rossini, Burrows-Wheeler transform. Bioinformatics 25, 1754–1760. doi: 10.1093/ L., et al. (2018). High-density multi-population consensus genetic linkage map bioinformatics/btp324 for peach. PloS One 13 (11), e0207724. doi: 10.1371/journal.pone.0207724 Li, X. W., Meng, X. Q., Jia, H. J., Yu, M. L., Ma, R. J., Wang, L. R., et al. (2013). Di Pierro, E. A., Gianfranceschi, L., Di Guardo, M., Koehorst-van Putten, H. J., Peach genetic resources: diversity, population structure and linkage Kruisselbrink, J. W., Longhi, S., et al. (2016). A high-density, multi-parental disequilibrium. BMC Genet. 14, 84. doi: 10.1186/1471-2156-14-84 SNP genetic map on apple validates a new mapping approach for outcrossing Li, H. (2011). A statistical framework for SNP calling, mutation discovery, association species. Hortic. Res. 3, 16057. doi: 10.1038/hortres.2016.57 mapping and population genetical parameter estimation from sequencing data. Dirks, R., van Dun, K., de Snoo, CB, van den Berg, M., Lelivelt, C. L. C., Voermans, Bioinformatics 27, 2987–2993. doi: 10.1093/bioinformatics/btr509 W., et al. (2009). Reverse breeding: a novel breeding approach based on Mercier, R., Mezard, C., Jenczewski, E., Macaisne, N., and Grelon, M. (2015). The engineered meiosis. Plant Biotechnol. J. 7, 837–845. doi: 10.1111/j.1467- Molecular Biology of Meiosis in Plants. Ann. Rev. Plant Biol. 66, 297–327. 7652.2009.00450.x Micheletti, D., Dettori, M. T., Micali, S., Aramini, V., Pacheco, I., Da Silva Donoso, J. M., Picañol, R., Eduardo, I., BatlIe, I., Howad, W., Aranzana, M. J., et al. Linge, C., et al. (2015). Whole-genome analysis of diversity and SNP-major (2015). High-density mapping suggests a cytoplasmic male sterility system gene association in peach germplasm. PloS One 10, e0136803. doi: 10.1371/ with two restorer factors in almond x peach progenies. Hortic. Res. 2, 15016. journal.pone.0136803 doi: 10.1038/hortres.2015.16 Sá nchez-Pé rez, R., Howad, W., Dicenta, F., Arú s, P., and Martı́nez-Gó mez, P. Doyle, J. J., and Doyle, J. L. (1990). Isolation of plant DNA from fresh tissue. Focus (2007). Mapping major genes and quantitative trait loci controlling agronomic 12, 13–15. traits in almond. Plant Breed. 125, 1–9. doi: 10.1111/j.1439-0523.2007.01329.x Frontiers in Plant Science | www.frontiersin.org 7 August 2020 | Volume 11 | Article 1205 Eduardo et al. Resynthesis Serra, O., Giné-Bordonaba, J., Eduardo, I., Bonany, J., Echeverrı́a, G., Larrigaudière, G., Ward, J. A., Bhangoo, J., Fernandez-Fernandez, F., Moore, P., Swanson, J. D., et al. (2017). Genetic analysis of the slow melting flesh character in peach. Tree Viola, R., et al. (2013). Saturated linkage map construction in Rubus idaeus Genet. Genomes 13, 77. doi: 10.1007/s11295-017-1160-x using genotyping by sequencing and genome-independent imputation. BMC Tello, J., Roux, C., Chouiki, H., Laucou, V., Sarah, G., Weber, A., et al. (2019). A Genomics 14, 2. doi: 10.1186/1471-2164-14-2 novel high-density grapevine (Vitis vinifera L.) integrated linkage map using Wijnker, E., Deurhof, L., van de Belt, J., de Snoo, C. B., Blankestijn, H., Becker, F., GBS in a half-diallel population. Theor. Appl. Genet. 132, 2237–2252. doi: et al. (2014). Hybrid recreation by reverse breeding in Arabidopsis thaliana. 10.1007/s00122-019-03351-y Nat. Protoc. 9, 761–772. doi: 10.1038/nprot.2014.049 Verde, I., Bassil, N., Scalabrin, S., Gilmore, B., Lawley, C. T., Gasic, K., et al. (2012). Development and evaluation of a 9k SNP array for peach by Conflict of Interest: The authors declare that the research was conducted in the internationally coordinated snp detection and validation in breeding absence of any commercial or financial relationships that could be construed as a germplasm. PloS One 7, e35668. doi: 10.1371/annotation/33f1ba92-c304- potential conflict of interest. 4757-91aa-555de64a0768 Verde, I., Abbott, A. G., Scalabrin, S., Jung, S., Shu, S., Marroni, F., et al. (2013). Copyright © 2020 Eduardo, Alegre, Alexiou and Arús. This is an open-access article The high-quality draft genome of peach (Prunus persica) identifies unique distributed under the terms of the Creative Commons Attribution License (CC BY). patterns of genetic diversity, domestication and genome evolution. Nat. Genet. The use, distribution or reproduction in other forums is permitted, provided the 45 (5), 487–494. doi: 10.1038/ng.2586 original author(s) and the copyright owner(s) are credited and that the original Voorrips, R. E. (2002). MapChart: Software for the graphical presentation of publication in this journal is cited, in accordance with accepted academic practice. No linkage maps and QTLs. J. Hered. 93, 77–78. doi: 10.1093/jhered/93.1.77 use, distribution or reproduction is permitted which does not comply with these terms. Frontiers in Plant Science | www.frontiersin.org 8 August 2020 | Volume 11 | Article 1205