Interspecific chromosomal effects on agronomic traits in Gossypium hirsutum by AD analysis using intermated G. barbadense chromosome substitution lines

The untapped potential of the beneficial alleles from Gossypium barbadense L. has not been well utilized in G. hirsutum L. (often referred to as Upland cotton) breeding programs. This is primarily due to genomic incompatibility and technical challenges associated with conventional methods of interspecific introgression. In this study, we used a hypoaneuploid-based chromosome substitution line as a means for systematically introgressing G. barbadense doubled-haploid line ‘3-79’ germplasm into a common Upland genetic background, inbred ‘Texas marker-1’ (‘TM-1’). We reported on the chromosomal effects, lint percentage, boll weight, seedcotton yield and lint yield in chromosome substitution CS-B (G. barbadense L.) lines. Using an additive-dominance genetic model, we studied the interaction of alleles located on two alien substituted chromosomes versus one alien substituted chromosome using a partial diallel mating design of selected CS-B lines (CS-B05sh, CS-B06, CS-B09, CS-B10, CS-B12, CS-B17 and CS-B18). Among these parents, CS-B09 and CS-B10 were reported for the first time. The donor parent 3-79, had the lowest additive effect for all of the agronomic traits. All of the CS-B lines had significant additive effects with boll weight and lint percentage. CS-B10 had the highest additive effects for lint percentage, and seedcotton and lint yield among all of the lines showing a transgressive genetic mode of inheritance for these traits. CS-B09 had greater additive genetic effects on lint yield, while CS-B06, CS-B10 and CS-B17 had superior additive genetic effects on both lint and seedcotton yield compared to TM-1 parent. The 3-79 line had the highest dominance effects for boll weight (0.513 g) and CS-B10 had the lowest dominance effect for boll weight (?0.702). Some major antagonistic genetic effects for the agronomic traits were present with most of the substituted chromosomes and chromosome arms, a finding suggested their recalcitrance to conventional breeding efforts. The results revealed that the substituted chromosomes and arms of 3-79 carried some cryptic beneficial alleles with potential to improve agronomic traits including yield, whose effects were masked at the whole genome level in 3-79.     

A Comparison of Leaf Anatomy in Field-grown Gossypium hirsutum and G. barbadense

Gossypium hirsutum L. (upland cotton) and G. barbadense L. (Pimacotton) are two of the most important fibre producing cottonspecies in cultivation. When grown side-by-side in the field,G.hirsutum has higher photosynthetic and transpiration rates (Luet al., 1997. Australian Journal of Plant Physiology24: 693–700).The present study was undertaken to determine if the differencesin physiology can be explained by leaf and canopy morphologyand anatomy. Scanning electron microscopy was used to comparethe leaf anatomy of field-grown upland (‘Delta’and ‘Pine Land 50’) and Pima (‘S6’)cotton. Compared to G. hirsutum, mature leaves of G. barbadenseare larger and thinner, with a thinner palisade layer. G. barbadenseleaves show significant cupping or curling which allows fora more even absorption of insolation over the course of theday and much more light penetration into the canopy. AlthoughG. barbadense leaves have a 70–78% higher stomatal densityon both the abaxial and the adaxial surfaces, its stomates areonly one third the size of those of G. hirsutum. This resultsin G. barbadense having only about 60% of the stomatal surfacearea per leaf surface area compared to G. hirsutum. These resultsare indicative of the anatomical and physiological differencesthat may limit the yield potential of G. barbadense in certaingrowing environments. Copyright 2000 Annals of Botany Company Cotton, leaf anatomy, leaf development, photosynthesis, Gossypium hirsutum, Gossypium barbadense, stomatal density     

Delineation of interspecific epistasis on fiber quality traits in Gossypium hirsutum by ADAA analysis of intermated G. barbadense chromosome substitution lines

Genetic diversity is the foundation of any crop improvement program, but the most cultivated Upland cotton [Gossypium hirsutum L., 2n?=?52, genomic formula?2(AD)₁] has a very narrow gene pool resulting from its evolutionary origin and domestication history. Cultivars of this cotton species (G. hirsutum L.) are prized for their combination of exceptional yield, other agronomic traits, and good fiber properties, whereas the other cultivated 52-chromosome species, G. barbadense L. [2n?=?52, genomic formula?2(AD)₂], is widely regarded as having the opposite attributes. It has exceptionally good fiber qualities, but generally lower yield and less desirable agronomic traits. Breeders have long aspired to combine the best attributes of G. hirsutum and G. barbadense, but have had limited success. F₁ hybrids are readily created and largely fertile, so the limited success may be due to cryptic biological and technical challenges associated with the conventional methods of interspecific introgression. We have developed a complementary alternative approach for introgression based on chromosome substitution line, followed by increasingly sophisticated genetic analyses of chromosome-derived families to describe the inheritance and breeding values of the chromosome substitution lines. Here, we analyze fiber quality traits of progeny families from a partial diallel crossing scheme among selected chromosome substitution lines (CS-B lines). The results provide a more detailed and precise QTL dissection of fiber traits, and an opportunity to examine allelic interaction effects between two substituted chromosomes versus one substituted chromosome. This approach creates new germplasm based on pair wise combinations of quasi-isogenic chromosome substitutions. The relative genetic simplicity of two-chromosome interactions departs significantly from complex or RIL-based populations, in which huge numbers of loci are segregating in all 26 chromosome pairs. Data were analyzed according to the ADAA genetic model, which revealed significant additive, dominance, and additive-by-additive epistasis effects on all of the fiber quality traits associated with the substituted chromosome or chromosome arm of CS-B lines. Fiber of line 3-79, the donor parent for the substituted chromosomes, had the highest Upper Half Mean length (UHM), uniformity ratio, strength, elongation, and lowest micronaire among all parents and hybrids. CS-B16 and CS-B25 had significant additive effects for all fiber traits. Assuming a uniform genetic background of the CS-B lines, the comparative analysis of the double-heterozygous hybrid combinations (CS-B?×?CS-B) versus their respective single heterozygous combinations (CS-B?×?TM-1) demonstrated that interspecific epistatic effects between the genes in the chromosomes played a major role in most of the fiber quality traits. Results showed that fiber of several hybrids including CS-B16?×?CS-B22Lo, CS-B16?×?CS-B25 and CS-B16?×?TM-1 had significantly greater dominance effects for elongation and hybrid CS-B16?×?CS-B17 had higher fiber strength than their parental lines. Multiple antagonistic genetic effects were also present for fiber quality traits associated with most of the substituted chromosomes and chromosome arms. Results from this study highlight the vital importance of epistasis in fiber quality traits and detected novel effects of some cryptic beneficial alleles affecting fiber quality on the 3-79 chromosomes, whose effects were not detected in the 3-79 parental lines.     

Molecular mechanisms of fiber differential development between G. barbadense and G. hirsutum revealed by genetical genomics

Cotton fiber qualities including length, strength and fineness are known to be controlled by genes affecting cell elongation and secondary cell wall (SCW) biosynthesis, but the molecular mechanisms that govern development of fiber traits are largely unknown. Here, we evaluated an interspecific backcrossed population from G. barbadense cv. Hai7124 and G. hirsutum acc. TM-1 for fiber characteristics in four-year environments under field conditions, and detected 12 quantitative trait loci (QTL) and QTL-by-environment interactions by multi-QTL joint analysis. Further analysis of fiber growth and gene expression between TM-1 and Hai7124 showed greater differences at 10 and 25 days post-anthesis (DPA). In this two period important for fiber performances, we integrated genome-wide expression profiling with linkage analysis using the same genetic materials and identified in total 916 expression QTL (eQTL) significantly (P<0.05) affecting the expression of 394 differential genes. Many positional cis-/trans-acting eQTL and eQTL hotspots were detected across the genome. By comparative mapping of eQTL and fiber QTL, a dataset of candidate genes affecting fiber qualities was generated. Real-time quantitative RT-PCR (qRT-PCR) analysis confirmed the major differential genes regulating fiber cell elongation or SCW synthesis. These data collectively support molecular mechanism for G. hirsutum and G. barbadense through differential gene regulation causing difference of fiber qualities. The down-regulated expression of abscisic acid (ABA) and ethylene signaling pathway genes and high-level and long-term expression of positive regulators including auxin and cell wall enzyme genes for fiber cell elongation at the fiber developmental transition stage may account for superior fiber qualities.     

Regeneration of Gossypium hirsutum and G. barbadense from shoot apex tissues for transformation

A method of regenerating cotton plants from the shoot apical meristem of seedlings was developed for use with particle gun and Agrobacterium-mediated transformation. This method was developed to circumvent the problems of genotype restriction and chromosomal damage frequently encountered in cotton regeneration in tissue culture through somatic embryogenesis. In this procedure, the cells of the shoot meristem are targeted for transformation. Normal and fertile plants of Gossypium barbadense Pima S-6, and 19 cultivars of G. hirsutum were regenerated using this method. Shoot regeneration from these tissues was direct and relatively rapid. A MS based, hormone-free medium could be used with all the varieties tested.This project was funded by grants from Cotton Incorporated, Nisshinbo Industries, and a grant from the Texas Agricultural Experiment Station to RHS. Texas Agricultural Experiment Station Technical Article TA-25667.     

Isolation,culture, and cell division in cotyledon protoplasts of cotton (Gossypium hirsutum and G. barbadense)

Protoplasts were isolated from cotyledons and foliage leaves of cotton (Gossypium hirsutum and G. barbadense). Cotyledon protoplasts were larger and responded to culture better than leaf protoplasts. Cotyledon derived protoplasts regenerated cell walls and formed microcolonies of 2–3 cells in G. hirsutum and 5–8 cells in G. barbadense. However, the microcolonies did not grow beyond this stage. Protoplast yield and viability, cell wall regeneration and cell division were influenced by several factors, e.g., genotype, age, tissue and growth condition of donor plant, enzyme mixture and concentration, preplasmolysis period, incubation period, and culture medium.Abbreviations 2,4-D 2,4-dichlorophenoxyacetic acid - NAA -naphthaleneacetic acid - BAP 6-benzylaminopurine - GA₃ gibberellic acid - p CPA p-chlorophenoxyacetic acid - MES 2[N-morpholino]ethanesulfonic acid     

Isolation, characterization and mapping of genes differentially expressed during fibre development between Gossypium hirsutum and G. barbadense by cDNA-SRAP

Gossypium hirsutum and G. barbadense are two cultivated tetraploid cotton species with differences in fibre quality. The fibre of G. barbadense is longer, stronger and finer than that of G. hirsutum. To isolate genes expressed differently between the two species during fibre development, cDNA-SRAP (sequence-related amplified polymorphism) was applied. This technique was used to analyse genes at different stages of fibre development in G. hirsutum cv. Emian22 and G. barbadense acc. 3-79, the parents of our interspecific mapping population. A total of 4096 SRAP primer combinations were used to screen polymorphism between the DNA of the parents, and 275 highly polymorphic primers were picked out to analyse DNA and RNA from leaves and fibres at different developmental stages of the parents. A total of 168 DNA fragments were isolated from gels and sequenced: 54, 30, 38 and 41 from fibres of 5, 10, 15 and 20 days post-anthesis, respectively, and five from multi stages. To genetically map these sequences, 104 sequence-specific primers were developed and were used to screened polymorphism between the mapping parents. Finally, six markers were mapped on six chromosomes of our backbone interspecific genetic map. This work can give us a primary knowledge of differences in mechanism of fibre development between G. hirsutum and G. barbadense.     

SSR marker-assisted improvement of fiber qualities in Gossypium hirsutum using G. barbadense introgression lines

Key message

This study demonstrates the first practical use of CSILs for the transfer of fiber quality QTLs into Upland cotton cultivars using SSR markers without detrimentally affecting desirable agronomic characteristics.

Abstract

Gossypium hirsutum is characterized by its high lint production and medium fiber quality compared to extra-long staple cotton G. barbadense. Transferring valuable traits or genes from G. barbadense into G. hirsutum is a promising but challenging approach through a traditional interspecific introgression strategy. We developed one set of chromosome segment introgression lines (CSILs), where TM-1, the genetic standard in G. hirsutum, was used as the recipient parent and the long staple cotton G. barbadense cv. Hai7124 was used as the donor parent by molecular marker-assisted selection (MAS). Among them, four CSILs, IL040-A4-1, IL080-D6-1, IL088-A7-3 and IL019-A2-6, found to be associated with superior fiber qualities including fiber length, strength and fineness QTL in Xinjiang were selected and backcrossed, and transferred these QTLs into three commercial Upland cotton cultivars such as Xinluzao (XLZ) 26, 41 and 42 grown in Xinjiang. By backcrossing and self-pollinating twice, five improved lines (3262-4, 3389-2, 3326-3, 3380-4 and 3426-5) were developed by MAS of background and introgressed segments. In diverse field trials, these QTLs consistently and significantly offered additive effects on the target phenotype. Furthermore, we also pyramided two segments from different CSILs (IL080-D6-1 and IL019-A2-6) into cultivar 0768 to accelerate breeding process purposefully with MAS. The improved lines pyramided by these two introgressed segments showed significant additive epistatic effects in four separate field trials. No significant alteration in yield components was observed in these modified lines. In summary, we first report that these CSILs have great potential to improve fiber qualities in Upland cotton MAS breeding programs.     

Inheritance of long staple fiber quality traits of Gossypium barbadense in G. hirsutum background using CSILs

Gossypium hirsutum is a high yield cotton species that exhibits only moderate performance in fiber qualities. A promising but challenging approach to improving its phenotypes is interspecific introgression, the transfer of valuable traits or genes from the germplasm of another species such as G. barbadense, an important cultivated extra long staple cotton species. One set of chromosome segment introgression lines (CSILs) was developed, where TM-1, the genetic standard in G. hirsutum, was used as the recipient parent and the long staple cotton G. barbadense Hai7124 was used as the donor parent by molecular marker-assisted selection (MAS) in BC₅S_1–4 and BC₄S_1–3 generations. After four rounds of MAS, the CSIL population was comprised of 174 lines containing 298 introgressed segments, of which 86 (49.4%) lines had single introgressed segments. The total introgressed segment length covered 2,948.7 cM with an average length of 16.7 cM and represented 83.3% of tetraploid cotton genome. The CSILs were highly varied in major fiber qualities. By integrated analysis of data collected in four environments, a total of 43 additive quantitative trait loci (QTL) and six epistatic QTL associated with fiber qualities were detected by QTL IciMapping 3.0 and multi-QTL joint analysis. Six stable QTL were detected in various environments. The CSILs developed and the analyses presented here will enhance the understanding of the genetics of fiber qualities in long staple G. barbadense and facilitate further molecular breeding to improve fiber quality in Upland cotton.     

Construction and characterization of the first bacterial artificial chromosome library for the cotton species Gossypium barbadense L.

As the second most widely cultivated cotton, Gossypium barbadense is well known for its superior fiber properties and its high levels of resistance to Fusarium and Verticillium wilts. To enhance our ability to exploit these properties in breeding programs, we constructed the first bacterial artificial chromosome (BAC) library for this species. The library contains 167 424 clones (49 920 BamHI and 117 504 HindIII clones), with an estimated average insert size of 130 kb. About 94.0% of the clones had inserts over 100 kb, and the empty clones accounted for less than 4.0%. Contamination of the library with chloroplast clones was very low (0.2%). Screening the library with locus-specific probes showed that BAC clones represent 6.5-fold genome equivalents. This high-quality library provides an additional asset with which to exploit genetic variation for cotton improvement.     

田间条件下海岛棉和陆地棉花铃期叶片光保护的机制

研究海岛棉(Gossypium barbadense)和陆地棉(G. hirsutum)两个棉花栽培种的光合作用特性, 探讨两个栽培种光合机构的光抑制以及防御保护机制, 以期为新疆棉花高光效品种选育和高产高效栽培实践提供理论基础。在新疆生态气候条件下, 系统测定了海岛棉和陆地棉的叶片运动、叶片接受光量子通量密度(PFD)、叶片温度、叶绿素荧光参数、气体交换参数和光呼吸速率的日变化。研究结果表明: 陆地棉叶片的“横向日性”较强而海岛棉较弱, 导致海岛棉叶片接受PFD较低, 但其叶片温度较高。海岛棉叶片的光合速率和气孔导度均显著低于陆地棉。在8:00-10:00 (北京时间, 下同)海岛棉叶片的光呼吸速率略低于陆地棉, 其余时间段海岛棉和陆地棉叶片的光呼吸速率相似。不同栽培种间, 叶片的最大光化学效率和实际光化学效率的日变化均无明显差异。除14:00-16:00以外, 海岛棉叶片的表观电子传递速率和光化学猝灭系数均显著低于陆地棉。8:00以后, 海岛棉叶片的非光化学猝灭显著高于陆地棉。因此, 在新疆生态气候条件下, 海岛棉和陆地棉叶片“横向日性”运动能力和气孔导度的差异导致叶片所处的光温环境不同, 同时造成海岛棉叶片的碳同化能力较低。为阻止光合电子传递链的过度还原, 减轻光合机构的光抑制, 陆地棉叶片主要通过光合机构的电子流途径耗散激发能, 而海岛棉叶片通过热耗散途径和相对较高的光呼吸能力来耗散激发能。     

Bidirectional cytoplasmic and nuclear introgression in the New World cottons Gossypium barbadense and G. Hirsutum (Malvaceae)

Patterns of interspecific cytoplasmic (plastid and mitochondrial) and nuclear introgression are typically asymmetrical: cytoplasmic gene flow or “capture” is frequently observed without evidence of nuclear introgression. In contrast, nuclear introgression without concomitant cytoplasmic introgression has rarely been demonstrated. Gossypium barbadense L. and G. hirsutum L. have large indigenous ranges in the New World semiarid tropics, with an extensive area of sympatry in the Caribbean and Central America. Numerous accessions of both species were surveyed for diagnostic cpDNA restriction sites. These data, in conjunction with previous information on nuclear markers, lead to several conclusions: 1) introgression between G. hirsutum and G. barbadense is bidirectional for both nuclear and cytoplasmic genes; 2) patterns of introgression between the two species are not symmetrical—in G. barbadense, introgression of G. hirsutum alleles is largely restricted to modem cultivars and is uncommon in areas of sympatry; in contrast, introgression of G. barbadense alleles into G. hirsutum is relatively common in areas of sympatry and is rare in modem cultivars; 3) nuclear introgression is geographically more widespread and more frequently detected than cytoplasmic introgression. Several mechanisms may underlie the differential patterns of cytoplasmic and nuclear gene flow observed, including differential fitness of infraspecific and interspecific cytonuclear combinations and selection against female function in interspecific backcrosses. Possible explanations for the observed asymmetrical patterns of introgression include differences in population sizes combined with phenological differences that bias interspecific pollen transfer.     

The distribution of Gossypium hirsutum chromatin in G. barbadense germ plasm: molecular analysis of introgressive plant breeding

Cotton is unusual among major crop plants in that two cross-fertile species are widely cultivated for a common economic product, fiber. Both historical evidence and classical genetic studies suggest that many improved forms of Gossypium barbadense (Sea Island, Egyptian, and Pima cottons) may include chromatin derived from G. hirsutum. Using 106 restriction fragment length polymorphism (RFLP) loci well distributed across the cotton genome, we revealed the amount and genomic distribution of G. hirsutum chromatin in 54 G. barbadense collections from around the world. The average G. barbadense collection was comprised of 8.9% alleles apparently derived from G. hirsutum. Pima cultivars (7.3 %) had fewer G. hirsutum alleles than Sea Island (9.0%) or Egyptian (9.6%) cultivars. G. hirsutum alleles were not randomly distributed, as 57.5% of the total introgression observed was accounted for by five specific chromosomal regions that span less than 10% of the genome. The average length of an introgressed chromosome segment was 12.9 cM. Overlap of introgressed chromatin in different breeding programs hints that retention of these G. hirsutum chromosomal segments may impart a selective advantage to G. barbadense genotypes. Although cluster analysis generally grouped germ plasm from common classes and/or breeding programs together, no 2 genotypes were identical — thus differences in the length and repertoire of introgressed chromosome segments also permit DNA fingerprinting of G. barbadense cultivars.     

Molecular dissection of interspecific variation between Gossypium hirsutum and G. barbadense (cotton) by a backcross-self approach: II. Fiber fineness

A backcross-self population from a cross between Gossypium hirsutum and G. barbadense was used to dissect the molecular basis of genetic variation governing two parameters reflecting lint fiber fineness and to compare the precision of these two measurements. By applying a detailed restriction fragment length polymorphism (RFLP) map to 3,662 BC₃F₂ plants from 24 independently derived BC₃ families, we were able to detect 32 and nine quantitative trait loci (QTLs) for fiber fineness and micronaire (MIC), respectively. The discovery of larger numbers of QTLs in this study than previously found in other studies based on F₂ populations grown in favorable environments reflects the ability of the backcross-self design to resolve smaller QTL effects. Although the two measurements differed dramatically in the number of QTLs detected, seven of the nine MIC QTLs were also associated with fiber fineness. This supports other data in suggesting that fiber fineness more accurately reflects the underlying physical properties of cotton fibers and, consequently, is a preferable trait for selection. Negative transgression, with the majority of BC₃F₂ families showing average phenotypes that were poorer than that of the inferior parent, suggests that many of the new gene combinations formed by interspecific hybridization are maladaptive and may contribute to the lack of progress in utilizing G. barbadense in conventional breeding programs to improve upland cotton.Electronic Supplementary Material Supplementary material is available for this article at     

Allozyme evidence for the origin and diversification of Gossypium barbadense L.

Summary Gossypium barbadense L. is a commercially important cotton species of tropical South American origin presently grownin many regions of the world. The species is morphologically diverse, consisting of a wide range of wild (or feral), commensal, landrace, and highly improvedcommercial forms. We performed allozyme analysis on 153 accessions representing the spectrum of G. barbadense diversityto ascertain the geographic origin of the species, its patterns of diffusion subsequent to domestication, and to reveal infraspecific relationships. Levels ofgenetic variation in G. barbadense are moderate. Of 59 loci scored, 24 were polymorphic, with a mean number of alleles perlocus of 1.69 and an average panmictic heterozygosity of 0.062. Principal component analysis revealed geographic clustering of accessions into six relativelydiscrete regions. Gene frequencies at many loci are significantly heterogeneous among these regions, with an average G _STof 0.272. Northwestern South America contains the greatest genetic variability; we suggest that this region is the ancestral home of the species. The data indicate separate diffusion pathways from this region into Argentina-Paraguay and into eastern and northern South America east of the Andes. Caribbean Island and Central American forms appear to be derived from the latter. These diffusion pathways are in accordance with morphological evidence and historical record. In contrast to expectations based on geographic proximity, Pacific Island forms have their closest affinity to accessions from eastern South America. Advanced cultivated stocks seem largely derived from western Andean material, but also contain introgressed G. hirsutum germ plasm. Introgression was relatively high (22%–50% of accessions) in commercial stocks and in forms from Argentina-Paraguay and various Pacific Islands, but was conspicuously low or absent in material from Central America and the Caribbean, where commensal and commercial forms of both species are sympatric.     

Enhanced unscheduled DNA synthesis in the cotyledons of Gossypium,barbadense L. by ethyl methanesulfonate (EMS)

When cotyledonary tissue of $\text{G.}\text{,}\text{barbadense}$ cotton are treated with the mutagen ethyl methanesulfonate and then germinated, an enhanced, unscheduled DNA synthesis response is observed, along with a concomitant increase in the thymidine triphosphate precursor pool size. The implications of these results are discussed in this paper.     

Molecular dissection of interspecific variation between Gossypium hirsutum and Gossypium barbadense (cotton) by a backcross-self approach: I. Fiber elongation

The current study is the first installment of an effort to explore the secondary gene pool for the enhancement of Upland cotton (Gossypium hirsutum L.) germplasm. We developed advanced-generation backcross populations by first crossing G. hirsutum cv. Tamcot 2111 and G. barbadense cv. Pima S6, then independently backcrossing F₁ plants to the G. hirsutum parent for three cycles. Genome-wide mapping revealed introgressed alleles at an average of 7.3% of loci in each BC₃F₁ plant, collectively representing G. barbadense introgression over about 70% of the genome. Twenty-four BC₃F₁ plants were selfed to generate 24 BC₃F₂ families of 22–172 plants per family (totaling 2,976 plants), which were field-tested for fiber elongation and genetically mapped. One-way analysis of variance detected 22 non-overlapping quantitative trail loci (QTLs) distributed over 15 different chromosomes. The percentage of variance explained by individual loci ranged from 8% to 28%. Although the G. barbadense parent has lower fiber elongation than the G. hirsutum parent, the G. barbadense allele contributed to increased fiber elongation at 64% of the QTLs. Two-way analysis of variance detected significant (P<0.001) among-family genotype effects and genotype×family interactions in two and eight regions, respectively, suggesting that the phenotypic effects of some introgressed chromosomal segments are dependent upon the presence/absence of other chromosomal segments.Electronic Supplementary Material Supplementary material is available for this article at