Introduction
It is now an accepted fact that domestication has had far-reaching effects on crop genomes. A common effect has been reduction in genetic diversity in crops compared with their wild progenitors because of population bottleneck or intensive selection (Tanksley and McCouch, Reference Tanksley and McCouch1997). Reduced diversity in crop cultivars is a growing concern because such crops lose wider adaptability and consistent productivity. Hence, there have been increased efforts to widen genetic base of the crops. Understanding the domestication genetics will greatly facilitate these efforts through discovery and utilization of rare but potentially important alleles present in these genetic resources.
Sesame (Sesamum indicum L.) is one of the oldest, most important and nutritious oilseed crops in the world. Sesame seed with high oil content along with antioxidants, vitamins and minerals makes it a desirable oilseed for consumption. Owing to its high economic importance, it is expected that the crop has been genetically modified in several regions of world, to improve its agronomic traits. This is evident from the improved traits of domesticated sesame, such as loss of seed dormancy, seeds being smooth and of varying colours, higher in oil content and large-sized capsules compared with its wild progenitors (Bedigian, Reference Bedigian2003). The history of domestication of sesame, particularly of Indian germplasm, has not been studied yet.
Domestication-associated genes offer an approach to reconstruct a crop's history of domestication using associated trait with phylogeny and phylogeographic resolution. No domestication-associated genes have yet been identified in sesame. Moreover, a variety of traits getting evolved during domestication make it more unclear as to what extent a single gene can be used to infer the cultivation history of whole crop genome.
Oil accounts for up to 50% of sesame seed (Nzikou et al., Reference Nzikou, Matos, Bouanga-Kalou, Ndangui, Pambou-Tobi, Kimbonguila and Desobry2009), making it one of the major determinants of both sesame seed yield and quality. Strong selection by humans on oil traits is thus expected during sesame domestication. It is likely to have occurred via the genes in the fatty acid biosynthetic pathway. Three key genes, found in this pathway, are known to play major roles in unsaturated fatty acid production: stearoyl acyl carrier protein desaturase 2 (fad2) and omega 3 fatty acid desaturase (o3fad). Sad is the plant enzyme that catalyses the conversion of stearic acid C18:0 into oleic acid C18:1. Fad2 encodes 18:1 desaturase and they are responsible for conversion of oleic C18:1 into linoleic acid C18:2. O3fad is responsible for conversion of linoleic acid C18:2 into linolenic acid C18:3. No study has been undertaken in sesame using these or any other genes to study domestication history in the world.
The aim was, therefore, to explore the domestication history of Indian sesame germplasm using domestication-associated genes. For this study, sad, fad2 and o3fad DNA sequence variations were evaluated within and among four populations of sesame germplasm: wild species, landraces, introgressed lines and improved cultivars. The objective was to assess how DNA sequence variation was altered through human activities of cultivation and intensive breeding approximately over the past 4000 years in India.
Materials and methods
Plant material
A total of 99 accessions of sesame were sampled from the core collection, based on morphological traits, maintained by the National Bureau of Plant Genetic Resources (NBPGR), New Delhi, India. Genotypes from different groups were selected to have a wide array of genetic diversity for this study. These included 70 accessions of endemic landraces, 20 improved Indian cultivars, four accessions of wild progenitor (S. malabaricum syn S. mulayanum Nair.), one accession of a wild species (S. radiatum Schum. et Thonn.) and four accessions of introgressed lines (Table S1, available online).
Fatty acid profile
A profile of the fatty acid composition of all the accessions was generated by gas chromatography. The procedure followed including oil extraction, fatty acid methyl esterification and fatty acid profiling was the same as reported earlier by us (Mondal et al., Reference Mondal, Bhat and Srivastava2010)
DNA extraction
DNA was extracted from the leaf tissue of a single, 2-month-old plant of each accession by the cetyl trimethyl ammonium bromide (CTAB) method (Saghai-Maroof et al., Reference Saghai-Maroof, Soliman, Jorgensen and Allard1984). DNA was diluted to yield 20 ng/μl concentration for further use.
Polymerase chain reaction and sequencing
Polymerase chain reaction (PCR) primers were designed using the Oligos software (Institute of Biotechnology, University of Helsinki, Finland). Coding regions of the three candidate genes were targeted for designing the primers. For amplifying the three genes, 12 different primer pairs were used, because the primers were designed to amplify with approximately 100 bp overlap between the neighbouring amplified regions for each gene. PCRs were conducted in 25 μl of solution, which included 1 × Taq buffer without MgCl2, 2.5 mm MgCl2, 1 U Taq DNA polymerase, 0.25 μm of each primer, 2 mm dNTP mix, 2 μl of 20 ng of genomic DNA and sterile deionized water. PCRs were conducted under the following conditions: denaturation at 94°C for 2 min, 30 cycles of denaturation at 94°C for 1 min, annealing at different temperatures depending upon T m (optimal melting temperature) of particular primer for 1 min, extension at 72°C for 1 min and final extension at 72°C for 10 min. The primers and the annealing temperature for PCRs are listed in Table S2 (available online).
PCR products were cleaned using Axygen commericial kits (Axygen Biosciences 33210 Central Avenue, Union City, CA 94587 USA), and then cycle sequenced using ABI Prism Big Dye Terminator cycle sequencer (Applied Biosystems, CA). In majority of the cases, PCR amplified products were sequenced from both directions to reduce the error rates. Sequences obtained were compared with the sesame genome using the BLASTN program and only the best hits were kept for further analyses. After these filtering steps, the consensus sequences were aligned with CLUSTAL W (Thompson et al., Reference Thompson, Higgins and Gibson1994) and the alignment was manually corrected with a sequence editor. Before the analysis, alignments were trimmed to exclude regions or sequences with missing or low-quality data.
Population genetic analysis
Population genetic analyses of the aligned sequences were performed using the program DnaSP 5.10 (Librado and Rozas, Reference Librado and Rozas2009). Nucleotide diversity, the average pairwise nucleotide divergence, of a multiple alignment was calculated as π (Tajima, Reference Tajima1983; Nei, Reference Nei1987), and the number of segregating sites as θ (Watterson, Reference Watterson1975). Several measures of sequence variation were obtained, and they were haplotype number, the signal of selection, i.e. deviation from neutrality (Tajima, Reference Tajima1989; Fu and Li, Reference Fu and Li1993), and the frequency of recombination, i.e. the minimum number of recombination events (Hudson and Kaplan, Reference Hudson and Kaplan1985). The positions of single-nucleotide polymorphisms (SNPs) and indels for each accession were generated. The strength of linkage disequilibrium (LD) was estimated using the Z nS statistic (Kelly, Reference Kelly1997), which is the average of R 2 (squared correlation coefficient) (Hill and Robertson, Reference Hill and Robertson1968) over all pairwise comparisons. DnaSP was also used to calculate Z a (Rozas et al., Reference Rozas, Gullaud, Blandin and Aguadé2001), the average of R 2 (Hill and Robertson, Reference Hill and Robertson1968) over all pairwise comparisons between adjacent polymorphic sites; ZZ (Rozas et al., Reference Rozas, Gullaud, Blandin and Aguadé2001) = Z a− Z nS; and Wall's B and Q tests (Wall, Reference Wall1999).
Sequences were collapsed into haplotypes and analysed using Arlequin version 3.5.1.2 (Excoffier and Lischer, Reference Excoffier and Lischer2010). Within- and between-population pairwise distances, population comparisons (F ST, 100 permutations) and molecular diversity (π, as a measure of nucleotide diversity, i.e. the probability that two randomly chosen homologous nucleotides are different) were computed by the pairwise distance method at a 0.05 significance level.
Results
Nucleotide diversity in sesame
A total of 4.4 Mbp of genomic DNA was sequenced, which comprised the coding region of the three unsaturated fatty acid-producing genes. In all the 99 sesame genotypes, 1501 bp of sad, 1437 bp of fad2 and 1470 bp of o3fad were targeted. A total of 716 variable sites were found (460 singleton and 240 parsimony informative sites). In addition, 578 indel sites were identified, which along with variable sites account for SNPs. The total number of mutations in these three genes was 514, of which 137 were synonymous changes and 277 were replacement changes.
To describe nucleotide diversity or sequence variation, the two common measures used were π, the expected heterozygosity per nucleotide site (Tajima, Reference Tajima1983), and θ, the number of segregating sites in a genotypic sample corrected for sample size (Watterson, Reference Watterson1975). The wild sesame had a substantial diversity with an average π value of 0.0116 and θ value of 0.0121, while the improved cultivars had comparatively reduced diversity (π = 0.0062 and θ = 0.0112). The landraces showed high diversity, while that of introgressed lines was quite low (Table S4, available online). Overall, improved cultivars retained 84.39% π and 37.45% θ of landraces (Table 1).
Table 1 Nucleotide diversity per base pair in coding regions within the four populations of sesame
Impr. cultivars, improved cultivars; π, the expected heterozygosity per nucleotide site; θ, the number of polymorphic sites in a genotypic sample corrected for sample size.
Domestication bottleneck
Domestication bottleneck represents the first occurrence of human-mediated selection (Hyten et al., Reference Hyten, Song, Zhu, Choi, Nelson, Costa and Specht2006). The nucleotide diversity retained by improved cultivars was 53.39% π and 92.56% θ of its wild population. The haplotype diversity (H d) data showed a higher genetic diversity in wild species for two loci, fad2 and o3fad (Table 2). Moreover, the matrix of pairwise F ST showed that wild-improved cultivar populations had high values (0.1–0.2) in all three loci, reflecting appreciable divergence between the two populations (Fig. 1, row 2). The neutrality tests (Tajima's D, Fu and Li's D, and Fu and Li's F) revealed a negative but not significant result in all the three loci for wild population (Table 3). Wild species showed a positive result of Fu's F test because of deficiency of alleles. This might be the result of population bottleneck or over dominant selection or small population size taken for the analysis. π non-synonymous (π non-syn) was also more in wild population in all the three loci when compared with improved cultivars, supporting the reduction in genetic diversity as expected (Table 4). The number of alleles was high in the improved cultivars when compared with the wild relatives in all three loci (Fig. 2). The values of LD, Z nS and Z a decreased in improved cultivars when compared with their wild forms. The ZZ value depicted a decrease only in o3fad locus and not in other two loci of improved cultivars compared with their wild relatives, probably because of the less number of accessions of wild forms taken for study. Moreover, Wall's B and Wall's Q reduced remarkably in value in case of improved cultivars against wild species (Table 2).
Table 2 Linkage disequilibrium and haplotype diversity (H d) within the four populations of sesame
Impr. cultivars, improved cultivars; R m, minimum number of recombination events in a sample; S, number of segregating sites; Z nS, average of R 2 over all pairwise comparisons; Z a, the average of R 2 over all pairwise comparisons between adjacent polymorphic sites; ZZ, Z a− Z nS; Wall's B, proportion of pairs of adjacent segregating sites that are congruent; Wall's Q, measure of linkage disequilibrium between adjacent pairs of segregating sites; H, Haplotype.
Fig. 1 Genetic diversity calculations for (a) Sad, (b) Fad2 and (c) O3fad in four populations of sesame. First row diagrams plot the average number of population pairwise differences (i.e. the mean number of basepair differences between all pairs of sequences in a sample): above diagonal shows the average number of pairwise differences between populations, diagonal element indicates the average number of pairwise differences within population and below diagonal is the corrected average pairwise difference (Nei's distance). Second row diagrams plot matrices of population pairwise F STs. A high F ST implies a considerable degree of differentiation among populations. Distance method: pairwise difference.
Table 3 Neutrality tests for sad, fad2 and o3fad loci within the four populations of sesame
Impr. cultivars, improved cultivars; n.d., not detectable.
Values were significantly different (*P< 0.05, **P< 0.01, ***P< 0.001).
Table 4 π non-synonymous (π non-syn) and π synonymous (π syn) of the three loci within four populations of sesame
Impr. cultivars, improved cultivars.
Fig. 2 A graphical presentation of variation in allele frequency at different loci in four populations of sesame using analysis of molecular variance (AMOVA).
The landraces had an average π value of 0.0074 and a θ value of 0.0299 (Table S4, available online) and were less diverse than the wild species, while it was more diverse than the improved cultivars. The landraces retained 63.26% (π) of diversity present in wild population. H d showed a retention of 63.81% in landraces from wild species (Table 2). Landraces had a high average number of pairwise differences with wild species in all three loci (Fig. 1, row 1). Moreover, o3fad locus showed that there is a very high reduction within population of landrace compared with that of wild. The F ST pairwise matrix of fad2 and o3fad loci showed a very high value (0.2–0.4) resulting in high divergence between the two populations (Fig. 1, row 2). The landraces showed a very high number of alleles compared with the wild species (Fig. 2). All the neutrality tests (Table 3) in landraces gave significantly negative results indicating an excess of rare polymorphism in the population, which is consistent with either positive selection or an increase in population size. The values of Z nS, ZZ and Z a decreased in landraces when compared with wild species. Wall's B and Wall's Q of landraces also showed less value than those of wild species (Table 2).
Effect of selection on diversity
Improved cultivar versus landrace germplasm
It is generally expected that intensive artificial selection imposed in modern plant breeding programs over the last century has reduced the genetic diversity from that present in the landraces. In this study, a comparison of the improved cultivars with the landraces indicated, on an average, a reduction in nucleotide diversity in improved cultivars for both π and θ, though not significantly (Table S4, available online).
Haplotype is a term used to designate a specific combination of linked alleles within a contiguous segment of DNA, and thus H d provides another measure for genetic diversity (Hyten et al., Reference Hyten, Song, Zhu, Choi, Nelson, Costa and Specht2006). The H d data of improved cultivars depicted a decrease in fad2 and sad loci when compared with those of landraces (Table 2). In improved cultivars' o3fad locus, there was an increase in the H d. These data were supported by the coalescent simulation, which was run 103 times, according to which, the H d in improved cultivars' sad and fad2 showed a decrease while that of o3fad showed an increase, in comparison to those of landrace population. The average number of pairwise difference between the two populations is very high in sad and fad2, while in o3fad it is moderately high (Fig. 1, row 1). Moreover, the number of alleles in all the three loci showed a higher value in landraces compared with improved cultivars (Fig. 2).
The pairwise neutrality tests, Tajima's D, Fu and Li's D, and Fu and Li's F data revealed a significant negative result in case of landraces for all the three loci (Table 3). The values ranged between − 2.4693 (sad) and − 2.7619 (fad2) for Tajima's D; − 4.2558 (sad) and − 5.9203 (fad2) for Fu and Li's D, and − 4.2569 (sad) and − 5.5486 (fad2) in case of Fu and Li's F. The significant negative values indicated an excess of rare polymorphism in the population, which might be consistent with either positive selection or an increase in population size. Improved cultivars also showed a significant negative result for o3fad locus, while fad2 and sad loci depicted a non-significant negative value. Fu's F analysis exhibited negative results for all the loci implying that there was an excess of alleles due to population expansion or genetic hitchhiking in both improved cultivars and landraces. The F ST pairwise matrix of all the three loci showed a low value (0.0–0.1) causing mild divergence between the two populations (Fig. 1, row 2).
Landrace versus introgressed germplasm
The results indicated a significant reduction in nucleotide diversity in introgressed lines when compared with that of landraces. On an average, introgressed lines retained only 17.77% (π) and 4.57% (θ) of landrace diversity, respectively (Table S4, available online). This could be supported by the H d of the three loci as all of them showed a decreased value in introgressed lines when compared with landraces (Table 2).
The neutrality tests, Tajima's D, Fu and Li's D, and Fu and Li's F values of landrace and introgressed lines population showed a significant difference in all three loci (Table 3). The data supported either a positive selection or an increase in population size in case of landrace. Fu's F value for all the three loci in landrace population provided negative results implying population expansion or genetic hitchhiking. The same could not be predicted for introgressed lines as fad2 and sad were undetectable and o3fad value of Fu's F was positive.
Selection in fatty acid desaturase loci
The genetic diversity in terms of π and θ was generally reduced in improved cultivars when compared with wild population. This reduction was recorded in two loci, o3fad and fad2, of improved cultivars, while sad locus depicted a considerable increase in these values (Table 1). There was also a decrease in H d in these two loci of improved cultivars (H d= 0.00 947 and 0.00 720, respectively) when compared with that of wild species (H d= 0.02 221 and 0.01 168 respectively) (Table 2). In case of sad locus, there was an increase in improved cultivars (H d= 0.00 181) than that of wild species (H d= 0.00 068). Moreover, in both o3fad and fad2 loci, the ratio of π non-synonymous:π synonymous (π non-syn:π syn) showed an increase in improved cultivars when compared with wild species. However, sad locus did not follow the trend as the ratio for wild was zero (Table 4). The neutrality tests of Tajima's D, Fu and Li's F, and Fu and Li's D presented negative values for all the three loci in wild as well as improved cultivars (Table 3). Comparison of these values in the two populations suggested that there was a remarkable decrease in improved cultivars compared with wild species. Only the o3fad gene of improved cultivars showed a significant decrease (away from neutral) in these values (Table 3). Even the Fu's Fs Stats showed a decrease in their values for the three loci in improved cultivars compared with wild species. Though not significant, the pairwise F ST matrix of improved cultivar, wild in all three loci, suggested high divergence between the two populations (Fig. 1, row 2).
Discussion
Nucleotide diversity in sesame
This study represents a comprehensive analysis of DNA sequence in germplasm of four diverse populations of Indian sesame germplasm, namely wild species, landraces, improved cultivars and introgressed germplasm. Although there was locus-to-locus variation in terms of π and θ values, it was apparent that wild sesame contains substantial levels of nucleotide diversity. Indeed, wild sesame appears to harbour ten times more diversity (π = 0.01 165; θ = 0.01 210) compared with that of many other crops including Glycine soja [π = 0.00 217 and θ = 0.00 235; (Hyten et al., Reference Hyten, Song, Zhu, Choi, Nelson, Costa and Specht2006)], wild Arabidopsis thaliana for PgiC locus [π = 0.003 and θ = 0.0065; (Kawabe et al., Reference Kawabe, Yamane and Miyashita2000)] and loblolly pine [π = 0.0040 and θ = 0.0041; (Brown et al., Reference Brown, Gill, Kuntz, Langley and Neale2004)]. By contrast, improved cultivated sesame contains markedly less nucleotide variation, which showed an average π value of 0.0062 and a θ value of 0.0112 (Table S4, available online). When compared with other organisms, the average π value of improved sesame cultivars was similar to that of flanking region of the sad2 gene of Linum usitatissimum [π = 0.00 650; (Fu, 2011)]. Moreover, the nucleotide diversity was similar to the θ values of GA20ox1 locus of Populus tremula [0.00 113 (Ingvarson, Reference Ingvarson2005)].
Domestication bottleneck
Crop cultivars are developed from wild progenitors through domestication and intensive selection for various features during breeding. It has been well documented that selection targeted at individual loci will reduce genetic diversity within and around the selected loci (Halliburton, Reference Halliburton2004). In this study, the domestication was supported by a reduction in genetic diversity of improved cultivars when compared with landraces (fad2 and sad) as well as wild population (o3fad and fad2). Overall, the effects of domestication combined with intensive selection in breeding have resulted in sequence diversity losses, in improved cultivars versus wild population, of 46.61% π, 7.44 % θ and 44.5% H d for all three loci, on an average. Similar results have been reported while studying domestication history of soybean (Hyten et al., Reference Hyten, Song, Zhu, Choi, Nelson, Costa and Specht2006), where the cultivars versus wild resulted in sequence diversity losses of 49% of π, 65% of θ and 44% of H d. Though the value of θ in this study did not show any significant loss of diversity, the rest of the values (including 39.10% loss of π non-syn) indicated a bottleneck in sesame domestication.
The landraces of sesame developed by farmers in different agroecological niche environment are long-run adaptations and selections in wild sesame species. These local cultigens were expected to have less genetic diversity during the course of time due to domestication. In this study, landraces retained 63.26% (π) of diversity present in wild progenitor. Similar results have been reported in soybean, in which there was 66% retention of π in landraces compared with the wild population (Hyten et al., Reference Hyten, Song, Zhu, Choi, Nelson, Costa and Specht2006). But sesame landraces showed an increase in θ when compared with that of wild relatives probably because the number of wild relatives taken for study is few compared with landraces. The H d value also suggested a loss of genetic diversity in landraces (0.0073) when compared with wild species (0.0115), which was 63.47% retention. This is in agreement with that of soybean where 63% retention of H d in landraces has been reported. Hence, the overall reduction in genetic diversity of landraces and improved cultivars suggested that genetic bottleneck had indeed occurred during the domestication of sesame due to the conscious and unconscious selection for oil quality in sesame by farmers over time.
It has been observed that the process of domestication leads to an increase in LD. But this study showed otherwise. Z nS, Z a and ZZ values decreased in improved cultivars and landraces when compared with the wild species. Same could be said about the values of Wall's B and Wall's Q. Main reason for decreased LD might be due to presence of frequent cross pollinations in landraces, which may result in recombinations between loci and thereby decreased LD. Further reason might be that sesame wild germplasm (seven species) in India is limited when compared with the 35 total species present in the Africas (Duhoon et al., Reference Duhoon, Sharma, Lakhanpaul, Bhat, Dhillon, Tyagi, Saxena and Agrawal2004). Thus, the genetic base for the domestication of sesame was very small leading to low LD in landraces and improved cultivars even after domestication.
The results of this study also suggested that there was no significant reduction in genetic diversity in improved cultivars compared with landraces. This was probably because involvement of hybridization was relatively less in frequency, during the development of improved cultivar from landraces. Majority of improved cultivars included in this study (12/20) were developed by direct selection from local collections/landraces, thereby resulting in direct representation of landrace diversity in released varieties (Table S1, available online).
There was very significant loss of diversity in the introgressed line population compared with all other populations included in this study. This might be explained by the fact that, through conventional method, introgressed line germplasm is developed during pre-breeding by hybridization of selected germplasm lines followed by intensive selection for trait(s). The method reduced genetic diversity significantly due to imposed selection pressure. Moreover, three out of four introgressed lines, included in this study, were the same across combination and the size of the introgressed line population was too small resulting in reduced genetic diversity as expected.
The results revealed that not all the genes depicted a similar diversity loss in improved cultivars compared to wild relatives and landraces. Particularly, o3fad showed higher diversity in improved cultivars than in landraces; sad showed very high diversity in improved cultivars. These results indicated that, during development of improved cultivars from landraces, selection was not targeted for this gene or its related genes. Similarly, in the second case (sad), there was either no or limited domestication or selection effect on the improved cultivars. Overall, results did highlight that despite domestication and selection, the Indian improved cultivars retained diversity in genes or traits to an extent suitable for cultivation in diverse agroclimatic regions of the country. This trend was also evident in earlier reports, such as the Indian cultivated forms that were rich in diversity (Duhoon et al., Reference Duhoon, Sharma, Lakhanpaul, Bhat, Dhillon, Tyagi, Saxena and Agrawal2004), and the collections showed higher molecular (random amplified polymorphic DNA; RAPD) diversity compared with that of exotic (Bhat et al., Reference Bhat, Babrekar and Lakhanpaul1999) varieties. One of the possible reasons for this trend could be that the majority of the Indian cultivars (37/62) were developed through selection from local collections/landraces (Duhoon et al., Reference Duhoon, Sharma, Lakhanpaul, Bhat, Dhillon, Tyagi, Saxena and Agrawal2004) to suit diverse local agroclimatic regions of the country.
Selection in fatty acid desaturase loci
This study included three desaturases, out of which two key genes involved were fad2 and o3fad, which catalyse the conversion of oleic acid into linoleic acid and linoleic into linolenic acid, respectively. The present results suggested that there was positive selection in them during domestication and/or improvement of cultivar in sesame. O3fad showed a reduction in genetic diversity as there was a loss of 57.44% of π, 11.67% of θ and 57.37% of H d in improved cultivars when compared with wild population. Similarly, in fad2 locus, there was a reduction in 38.36% π, 16.00% θ and 38.47% H d in improved cultivars. Earlier, Maynard and Haigh (Reference Maynard and Haigh1974) stated that a strong positive selection leads to a loss of nucleotide diversity.
The ratio of π non-syn:π syn was 1.021 for o3fad and 1.753 for fad2, which further supported the theory of positive selection for the two loci during the course of domestication. Similar results have been recorded in ods locus of Drosophila (Ting et al., Reference Ting, Tsaur, Wu and Wu1998), het-c of Sordariaceae fungi (Wu et al., Reference Wu, Saupe and Glass1998), S locus of various plants (Clark and Kao, Reference Clark and Kao1991), 19 seminal protein genes of Drosophila (Swanson et al., Reference Swanson, Clark, Waldrip-Dail, Wolfner and Aquadro2001) and RPP1complex, RPP8 complex, RPP13 of Arabidopsis thaliana (Botella et al., Reference Botella, Parker, Frost, Bittner-Eddy, Beynon, Daniels, Holub and Jones1998, McDowell et al., Reference McDowell, Dhandaydham, Long, Aarts, Goff, Holub and Dangl1998, Bittner-Eddy et al., Reference Bittner-Eddy, Crute, Holub and Beynon2000, respectively) and in several others. These studies lead to direct postulation of positive selection among these genes because their ratio of π non-syn:π syn was more than 1, which further supports the present data. Moreover, the matrix of pairwise F ST for all three loci in improved cultivar and wild species showed a significant elevation from zero. This suggested high divergence between two populations (Fig. 1, row 2).
Tajima's D, Fu and Li'D, and Fu and Li's F, along with Fu's F stats, were performed for studying neutrality in detail. Tajima's D, Fu and Li's D, and Fu and Li's F gave negative result for the improved cultivars of both o3fad and fad2, suggesting an excess of rare variant relative to neutral equilibrium model, and was thus consistent with a positive selection. But the values were significant only for improved cultivars of o3fad. This could be explained with the fact that demography was not considered in these loci. The population structure especially variation in the number of samples in different populations also affected the results.
The results suggested that there was a positive selection in o3fad and fad2. The phenotypic data also supported this result because improved cultivars showed an increase in both linoleic and linolenic acid contents (%) compared with wild species (Table S3, available online).
The observed loss of diversity from wild to improved cultivars was likely to, at least in part, cause population bottlenecks during the domestication of sesame. Despite domestication, the improved cultivated sesame showed variable results with different genes in this study. This can be supported by the fact that sesame has been considered as a semi-domesticated crop having still few important characters of wild species (Bedigian, Reference Bedigian2003).
Plant domestication, breeding and biotechnology have modified sesame genomes according to the needs of humanity with increasing efficiency and precision. Understanding such processes, crop domestication in particular, is crucial today because of the rising demand for improving yield and quality, as well as to obtain sesame with additional capability to grow in varied ecosystems. The background information obtained from this study indicated that, during crop improvement, selective genes were to be considered based on the domestication dynamics. Moreover, an increased number of wild species from around the world is needed to be included for further study to enhance the possibility of incorporation of desirable genes in cultivars from wild relatives.
Supplementary material
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S1479262115000106
Acknowledgements
The authors thank Dr P. Dhureja, Agricultural Chemicals, IARI, for laboratory facilities and Mr R. Tanwar for technical assistance in performing gas chromatography for this study. Financial assistance to the senior author in the form of Junior and Senior Research Fellowship by the University Grant Commission, Government of India, is duly acknowledged. The authors also thank NBPGR and National Agricultural Innovation Project for supporting this work and paper.
