The genetic diversity, relationships, and evolutionary history of the Gossypium species, including upland cotton (Gossypium hirsutum L.), island cotton (Gossypium barbadense L.), herbaceous cotton (Gossypium herbaceum L.), tree cotton (Gossypium arboreum L.), have been subjects of extensive study. These studies have been instrumental in providing critical insights that are essential for the genetic improvement of cotton. Recent advances such as the assembly of the Gossypium herbaceum L. genome and the enhanced genomes of Gossypium arboreum L. and Gossypium hirsutum L. have clarified the phylogenetic relationships and origins of cotton A-genomes. It is suggested that all existing A-genomes might have descended from a common ancestor, designated A0, which shows a closer relation to A1 than to A2 (Huang et al., 2020). This ancestor likely existed before the formation of allotetraploids, which speciated into A1 and A2 and subsequently evolved independently without a direct ancestor-progeny relationship (Huang et al., 2020).

The genetic diversity and relationships within the Gossypium taxa have also been examined using AFLP and RAPD markers. These studies have unveiled a broad spectrum of genetic similarities and robustly supported phylogenetic relationships (Abdalla et al., 2001; Khan et al., 2000). Notably, Gossypium arboreum L. and Gossypium raimondii Ulbr. were found to be genetically distinct, whereas Gossypium barbadense L. and Gossypium hirsutum L. demonstrated a high degree of genetic similarity (Abdalla et al., 2001). Moreover, the molecular phylogeny of Gossypium species, as revealed through DNA fingerprinting, has identified six primary clusters showing varied genetic similarities and has emphasized the genetic connections of Gossypium hirsutum L. with other cultivated species (Khan et al., 2000).

Further, the divergence between Gossypium hirsutum L. and Gossypium barbadense L. has been detailed through genome-anchored SNPs, which indicated minimal allele sharing between the two species, suggesting limited gene flow (Reddy et al., 2017). The repeated phenomenon of polyploidization in the Gossypium genomes has played a significant role in the evolution of spinnable cotton fibers. Allopolyploidy, in particular, has merged divergent genomes, leading to substantial duplication of ancestral genes (Paterson et al., 2012).

Phylogenetic diversity and relationships among Gossypium germplasm have been assessed using SSR markers, confirming that Gossypium raimondii Ulbr. is the closest living relative of the ancestral D-genome donor of tetraploid species. The A-genome donor exhibits similarity to present-day Gossypium herbaceum L. and Gossypium arboreum L. (Wu et al., 2007). These findings are crucial for understanding the genetic diversity and evolutionary processes that have shaped the Gossypium genus, thus providing a solid foundation for further research and breeding programs aimed at enhancing cotton production.

Gossypium remains a vital raw material for the global textile industry and holds significant economic value. The distinct types of cotton and their genetic traits directly influence the quality and yield of cotton fibers. Thus, research into genetic diversity not only enhances our scientific understanding of their evolutionary paths but also supports practical applications in improving cotton varieties and boosting agricultural productivity. The RAPD-PCR method, known for its simplicity, speed, and cost-effectiveness, is extensively utilized in plant genetic diversity and phylogenetics research. This technique facilitates the generation of numerous genetic markers to estimate genetic distances and construct phylogenetic relationships both within and between species. Understanding the genetic differences among various cotton species and their varieties is crucial for developing adaptable plants for future agricultural systems, particularly those facing challenges from climate change and pests.

In this context, our study utilized RAPD-PCR technology to analyze the DNA polymorphism of 21 cotton samples from different species, including upland cotton (Gossypium hirsutum L.), island cotton (Gossypium barbadense L.), herbaceous cotton (Gossypium herbaceum L.), and tree cotton (Gossypium arboreum L.). The aim was to delve into the genetic similarities, kinship, and evolutionary trends among these species and their various varieties, with the ultimate goal of leveraging this genetic information to guide cotton cultivation and variety improvement initiatives, thus promoting the advancement of the global cotton industry.

1 Materials and Methods

1.1 Cotton species tested

This study utilized cotton varieties from different genomic backgrounds (Table 1). The specific materials included 16 samples of Gossypium hirsutum L., three samples of Gossypium barbadense L., one sample of Gossypium herbaceum L., and one sample of Gossypium arboreum L. These were used to analyze genetic similarities, phylogenetic relationships, and evolutionary patterns among different cotton species and among different varieties within the same species.

1.2 Extraction of genomic DNA from cotton

Approximately 2-3 grams of fresh young leaves from the top of the cotton plant were ground into powder in liquid nitrogen. The powdered leaves were then transferred to a 15 mL centrifuge tube. Next, 6 mL of preheated cotton DNA extraction buffer at 65 °C was added to the tube. The mixture was incubated at 65 ℃ for 30 minutes to 1 hour to facilitate cell lysis. Subsequently, 6 mL of chloroform-isoamyl alcohol solution (24:1 ratio) was added. After thorough mixing, the mixture was centrifuged at 4 000 rpm for 20 minutes. This extraction with chloroform-isoamyl alcohol was repeated 2 to 3 times to completely separate the DNA. Afterwards, 4 mL of isopropanol was added, mixed well, and allowed to stand at room temperature for 5 minutes to precipitate the DNA. The flocculent DNA precipitate was then retrieved using a fine needle and washed 2 to 3 times with 75% ethanol to remove residual impurities. The precipitate was air-dried at room temperature or under a laminar flow hood. Finally, the DNA precipitate was redissolved in 500 µL of TE buffer (pH 8.0) for subsequent use.

The composition of the cotton DNA extraction buffer included: 0.1 M Tris-HCl (pH 8.0), 1.0 M NaCl, 0.02 M EDTA (pH 8.0), 2% (w/v) CTAB, 2% (w/v) PVP40, 1 mM 1,10-phenanthroline, and 0.2% (v/v) β-mercaptoethanol.

1.3 RAPD-PCR method

The RAPD-PCR analysis was carried out in a 15 μL PCR reaction mixture that included 1.5 μL of 10×PCR Buffer (500 mmol/L KCl; 100 mM Tris-HCl, pH 9.0; 1% Triton X-100), 1.2 μL of 20 mM MgCl₂, and 0.12 μL of 10 mM dNTP mix (each dATP, dCTP, dTTP, dGTP at 2.5 mM). Additionally, 1.6 μL of 100 pM RAPD primer (10-base random primers produced by the University of British Columbia, Canada) and 0.12 μL of 1.5 U Taq polymerase (from Promega) were added. The template DNA was added at 30 ng, and the volume was completed to 15 μL with ddH₂O.

PCR reactions were performed on a PTC-100 thermal cycler (MJ Research Inc.). The thermal cycling conditions were as follows: an initial denaturation step at 94 °C for 5 minutes, followed by 41 cycles of 1 minute at 94 °C for denaturation, 1.5 minutes at 36 °C for annealing, and 2 minutes at 72 °C for extension. A final extension step at 72 °C for 10 minutes was included to ensure complete reaction completion.

Following PCR amplification, the products were analyzed using polyacrylamide gel electrophoresis and silver staining. First, glass plates were thoroughly cleaned and rinsed with distilled water and allowed to dry before assembly. The gel was prepared by combining 4.5 mL of 30% polyacrylamide, 19.5 mL of ddH₂O, 6 mL of 5×TBE, 0.24 mL of 10% ammonium persulfate, and 0.014 mL of TEMED. After mixing evenly, the solution was poured slowly along a glass rod into the assembled glass plates, and a comb was inserted. The gel was left to polymerize for 30 minutes.

Electrophoresis was conducted at a constant voltage of 300 V for approximately 2 hours. Subsequently, the gel was subjected to rapid silver staining and band visualization, which included fixing (100 mL 10% ethanol+500 μL acetic acid, fixed for 2~3 mins), staining (addition of 1 mL 20% AgNO₃, stained for 5~8 minutes), washing (rinsed 2~3 times with distilled water), and developing (100 mL 10% NaOH+500 μL formaldehyde added). Finally, the gel was preserved and the band data were analyzed statistically.

1.4 Data recording and analysis

To record the banding data produced by RAPD-PCR, a binary encoding method was employed: samples showing bands at specific loci were recorded as “1”, and those without bands were recorded as “0”. This method facilitates subsequent data processing and analysis, and assists in calculating the polymorphism rate of RAPD markers, i.e., the frequency of band presence or absence among different samples.

In terms of data analysis, Clustal W software was used to process and align the banding data. A phylogenetic tree and corresponding genetic distance matrix based on banding differences were generated using the draw N-J tree function in the tree command menu of the software. This step primarily serves to display the genetic similarities and degrees of differentiation among different cotton samples.

For a more intuitive interpretation of the analysis results, DNAman software was utilized to analyze and edit the generated dendrograms. This software provides powerful visual tools that enable researchers to clearly identify the branching structures of genetic relationships, thereby offering deeper insights into the genetic connections and evolutionary history between different cotton varieties and species.

2 Results and Analysis

2.1 RAPD-PCR analysis of cotton materials tested

In this study, 14 random primers were used to perform RAPD-PCR analysis on 21 different cotton samples. Through these primers, a total of 649 alleles were detected, of which 116 exhibited polymorphism, resulting in a polymorphism rate of 17.9%. The number of alleles amplified by each primer ranged from 4 to 97, with an average of 41 alleles. The number of effective alleles amplified per primer varied from 1 to 13, with an average of 8.3.

Figure 1 displays the electrophoretic banding pattern obtained after RAPD-PCR using primer RA239 across all tested cotton materials. This figure clearly reveals the genetic differences among the cotton samples, where the presence and absence of polymorphic bands provide a crucial basis for further exploration of genetic diversity and the phylogenetic relationships within and between species.

2.2 Genetic distance analysis of cotton materials

Genetic distances were estimated using Kimura's parameter method (Kimura, 1980), which allows us to quantify genetic differences among cotton samples.

2.2.1 Genetic distance and cluster analysis of island cotton

As shown in Table 2, among the island cotton samples tested, the smallest genetic distance was between Hai 7124 and Ashmor, at 0.296, while the largest was between Hai 7124 and VH8-4602, at 0.630. The genetic distance between Ashmor and VH8-4602 was 0.519. These data reveal the relatively small genetic differences and closer phylogenetic relationship between Hai 7124 and Ashmor, while the larger genetic distance between VH8-4602 and Hai 7124 indicates a more distant phylogenetic relationship.

Cluster analysis of the three island cotton samples was conducted using the Neighbor-Joining (NJ) method. As depicted in Figure 2, the results categorized these samples into two main groups. Hai 7124 and Ashmor formed one group, indicating a closer genetic linkage between them. In contrast, VH8-4602 constituted a separate group, demonstrating a greater genetic distance from the other two island cottn varieties.

2.2.2 Genetic distance in upland cotton

According to the data in Table 3, among the 11 upland cotton samples, the smallest genetic distance was between Guangzimian and Zhong 19, at only 0.148, while the largest was between Ji 668 and Zhong 19, reaching 0.648. The genetic distances among these samples generally ranged from 0.148 to 0.648, with an average genetic distance of 0.366. This indicates significant genetic diversity among the Upland cotton samples, with a variation range significantly greater than that of the island cotton samples (0.296-0.630).

The clustering results for the upland cotton varieties are shown in Figure 3, where the 11 upland cotton varieties are divided into two major groups. The first group includes Zhong 19, Guangzimian, Xinluzao 7, 99B, Zhong 12, Kezimian, and Junmian 1. Within this group. Junmian 1 forms a separate sub-group, while another sub-group is further divided into two smaller groups: Zhong 12 and Kezimian cluster together first, and Zhong 19, Guangzimian, Xinluzao 7, and 99B form another group, with Zhong 19 and Guangzimian showing the closest phylogenetic relationship, and 99B being relatively more distantly related to the other varieties. The second group consists of Ji 668, 86-1, Liao 15, and ZYS22, with Ji 668 and 86-1 clustering first, indicating a more distant genetic relationship with Liao 15 and ZYS22.

2.2.3 Genetic distance among colored cotton samples

Among the four colored cotton samples tested, the genetic distance between DW1987 and AFSPT586 was 0.389, while the largest genetic distance was between Caimian 1 and IRG1919, at 0.574. The range of genetic distances among the other colored cotton samples also fell between 0.389 and 0.574, a range similar to that observed in upland cotton. Based on these genetic distances, appropriate colored cotton can be selected as parental lines for breeding new varieties with desirable traits (Table 4).

As shown in Figure 4, among the four colored cottons, Caimian 1 exhibits a more distant phylogenetic relationship with the other three varieties. DW1987 and IRG1919 cluster first, indicating a closer genetic relationship between them, while AFSPT586 is relatively closer to IRG1919 and DW1987, and more distantly related to Caimian 1. These clustering results highlight the effectiveness of RAPD markers in differentiating between different cotton materials and revealing their phylogenetic relationships.

Figure 4 NJ clustering diagram of RAPD markers for tested colored cottons

2.3 Genetic distance analysis among tested cotton materials and across four major cotton types

To further explore the phylogenetic relationships among 21 cotton samples, we calculated the genetic distances including all tested materials (Table 5). The range of genetic variation was 0.148 to 0.648, which is consistent with the categorized results. Detailed analysis revealed that the average genetic distance among island cotton samples was 0.241, among upland cotton samples was 0.366, among colored cotton samples was 0.281, and the average genetic distance across all 21 samples was 0.403. The average genetic distance between island cotton and upland cotton was 0.48, between island cotton and colored cotton was 0.478, and between colored cotton and other upland cottons was 0.430, all of which are higher than the overall average. This indicates that the phylogenetic relationships between different cotton types are relatively distant, while relationships within the same cotton type are closer, thereby validating the suitability of genetic distance analysis for studying the evolution and differentiation of cotton material resources.

Further, we conducted a detailed comparison of genetic distances among 21 cotton samples (Table 6). The results show that the average genetic distance between tree cotton and tetraploid cotton species (including island cotton and upland cotton) was 0.400, while the average genetic distance between herbaceous cotton and tetraploid cottons was 0.424. Overall, tree cotton has a closer phylogenetic relationship with tetraploid cotton species. Comparing the genetic distances between tree cotton, herbaceous cotton, and the island and upland cottons, tree cotton shows a smaller genetic distance to upland cotton, indicating a closer phylogenetic relationship; whereas herbaceous cotton is genetically closer to island cotton. Additionally, specific varieties such as Hai 7124 and VH8-4602 have a closer relationship with herbaceous cotton, while the Ashmor variety shows a closer genetic distance to tree cotton.

Table 6 Comparison of genetic distance between diploid and tetraploid cotton

2.4 Cluster analysis of cotton samples and the distribution of diploid cottons

Cluster analysis was performed on 21 cotton samples, including island cotton, upland cotton, colored cotton, and diploid cottons, using 116 marker data generated by 14 RAPD primers. The results are displayed in a dendrogram (Figure 5), which divided the 21 samples into two major groups.

Figure 5 NJ clustering diagram of RAPD markers for 21 cotton samples

The first group was further divided into two subgroups: The first subgroup includes Zhong 19, Guangzimian, AFSPT586, herbaceous cotton, Zhong 12, DW1987, FLS2123, 99B, Xinluzao 7, and Shixiya 1. Zhong 19 and Guangzimian first formed a closely related cluster, followed by clustering with AFSPT586 and herbaceous cotton, indicating close genetic links among them. Zhong 12 and DW1987 initially clustered together, forming a subgroup with FLS2123, while 99B, Xinluzao 7, and Shixiya 1 formed another group within the second subgroup. The second subgroup includes Kezimian, IRG1919, Junmian 1, and VH8-4602, with Kezimian and IRG1919 clustering first.

The second group includes Caimian 1 and 86-1 forming a subgroup with close phylogenetic relationships, and showing close genetic links with Ji 668. In contrast, Liao 15, ZYS22, as well as Hai 7124 and Ashmor exhibit more distant phylogenetic relationships.

From the dendrogram, it can be observed that the island cotton samples did not cluster together, nor did the upland cotton samples cluster with other upland cotton samples. This finding is inconsistent with traditional classification methods, and possible explanations include:

(1) Many of the samples come from cultivated varieties used in production, such as Ji 668, which possesses genetic characteristics of both island and upland cotton.

(2) The diversity of origins leads to a rich genetic diversity among the samples.

(3) During the long history of natural and artificial selection in cotton, there may have been gene flow between different chromosomal groups.

(4) The natural evolution and artificial breeding strategies of cotton have influenced its genetic structure.

3 Discussion

Genetic distance is a crucial quantitative indicator used to measure the degree of differentiation and genetic diversity between populations. In evolutionary biology, a larger genetic distance usually implies a more distant kinship and significant evolutionary divergence. In this study, the diversity analysis of 21 cotton germplasm resources revealed that the genetic distances between species (for example, the average genetic distance between island cotton and upland cotton is 0.48) are generally greater than the genetic distances within species (0.366 within upland cotton), indicating clear evolutionary divergence between species. Notably, the genetic distances between herbaceous cotton and tetraploid cotton (0.424) and between tree cotton and tetraploid cotton (0.400) suggest a closer kinship between tree cotton and tetraploid cotton, hinting at the diploid origins of the A genome in tetraploid cotton.

However, the genetic distance between tree cotton and island cotton (0.4317) is slightly greater than between herbaceous cotton and island cotton (0.4193), while the distance between tree cotton and upland cotton (0.3946) indicates a closer kinship between them. These results suggest that the A genomes of island cotton and upland cotton might originate from different diploid ancestors, with herbaceous cotton having a closer diploid ancestor relationship with island cotton, and tree cotton being more closely related to the diploid ancestors of upland cotton. This finding differs from some previous research results, indicating that the evolution of the cotton genome is complex and diverse.

Based on the above research evidence, we can discuss the genetic distance and its implications for evolutionary divergence within the genus Gossypium. The studies on Gossypium mitochondrial genomes reveal significant structural differences between diploid and allotetraploid species, indicating a rapid evolutionary divergence. For instance, the mitochondrial genomes of diploid species from the D group were found to be identical but differed from those of the A group and the AD group species, suggesting a genetic distance that reflects their evolutionary paths (Chen et al., 2017).

Furthermore, introgression events, such as those from Gossypium hirsutum to Gossypium barbadense, have been shown to contribute to population divergence and genetic diversity. This indicates that genetic distance can be increased through introgression, which introduces new genetic material and can drive evolutionary change (Wang et al., 2022).

Chloroplast DNA analysis among diploid cotton species has also provided insights into the phylogeny of Gossypium, with nucleotide distances being smaller within genome groups than among them. This supports the idea that genetic distance correlates with evolutionary divergence, as more closely related species within a group share more similar chloroplast DNA sequences (Chen et al., 2016).

In addition, a study on low-copy DNA sequence divergence among Gossypium genomes has revealed varying levels of polymorphism, with the A genome lineage evolving rapidly and the D genome showing accelerated evolution compared to the C genome. This suggests that genetic distance can vary even within closely related genome types, reflecting different rates of evolutionary change (Rong et al., 2012).

Lastly, an unusual ribosomal DNA sequence from Gossypium gossypioides has provided evidence of ancient, cryptic introgression, which has implications for understanding the genetic distance and evolutionary history of the species. This introgression event has likely contributed to the genetic divergence of Gossypium gossypioides from other American D-genome species (Wendel et al., 2015).

In summary, the genetic distance within the Gossypium is a reflection of the degree of differentiation and evolutionary divergence between species. The studies cited here demonstrate that genetic distance can be influenced by factors such as structural genome changes, introgression events, and varying rates of evolution among different genome types. These factors contribute to the complex evolutionary relationships within Gossypium, highlighting the importance of genetic distance as a quantitative indicator in evolutionary biology.

Acknowledgments

I would like to express my gratitude to my graduate advisor for their profound insights and valuable suggestions on the manuscript, which significantly improved the quality of this paper. This research was conducted at the Hainan Provincial Key Laboratory of Crop Molecular Breeding, and I am thankful for the experimental conditions and resources provided by the laboratory. I would also like to thank my laboratory colleagues for their support and assistance; their collaboration made it possible for my research to proceed smoothly. Furthermore, I extend my appreciation to the anonymous peer reviewers whose incisive and constructive comments enriched and enhanced the content of this study, adding to its constructiveness and scientific value.

Conflict of Interest Disclosure

The authors affirm that this research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.   

References

Abdalla A., Reddy O., El-Zik K., and Pepper A., 2001, Genetic diversity and relationships of diploid and tetraploid cottons revealed using AFLP, Theoretical and Applied Genetics, 102: 222-229.

https://doi.org/10.1007/s001220051639

Chen Z., Feng K., Grover C., Li P., Liu F., Wang Y., Xu Q., Shang M., Zhou Z., Cai X., Wang X., Wendel J., Wang K., and Hua J., 2016, Chloroplast DNA structural variation, phylogeny, and age of divergence among diploid cotton species, PLoS One, 11(6): e0157183.

https://doi.org/10.1371/journal.pone.0157183

PMid:27309527 PMCid:PMC4911064

Chen Z., Nie H., Wang Y., Pei H., Li S., Zhang L., and Hua J., 2017, Rapid evolutionary divergence of diploid and allotetraploid Gossypium mitochondrial genomes, BMC Genomics, 18(1): 876.

https://doi.org/10.1186/s12864-017-4282-5

PMid:29132310 PMCid:PMC5683544

Huang G., Wu Z., Percy R., Bai M., Li Y., Frelichowski J., Hu J., Wang K., Yu J., and Zhu Y., 2020, Genome sequence of Gossypium herbaceum and genome updates of Gossypium arboreum and Gossypium hirsutum provide insights into cotton A-genome evolution, Nature Genetics, 52: 516-524.

https://doi.org/10.1038/s41588-020-0607-4

PMid:32284579 PMCid:PMC7203013

Khan S., Hussain D., Askari E., Stewart J., Malik K., and Zafar Y., 2000, Molecular phylogeny of Gossypium species by DNA fingerprinting, Theoretical and Applied Genetics, 101: 931-938.

https://doi.org/10.1007/s001220051564

Kimura M., 1980, A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences, J. Mol. Evol., 16: 111-120.

https://doi.org/10.1007/BF01731581

PMid:7463489

Paterson A., Wendel J., Gundlach H., Guo H., Jenkins J., Jin D., Llewellyn D., Showmaker K., Shu S., Udall J., Yoo M., Byers R., Chen W., Doron-Faigenboim A., Duke M., Gong L., Grimwood J., Grover C., Grupp K., Hu G., Lee T., Li J., Lin L., Liu T., Marler B., Page J., Roberts A., Romanel E., Sanders W., Szadkowski E., Tan X., Tang H., Xu C., Wang J., Wang Z., Zhang D., Zhang L., Ashrafi H., Bedon F., Bowers J., Brubaker C., Chee P., Das S., Gingle A., Haigler C., Harker D., Hoffmann L., Hovav R., Jones D., Lemke C., Mansoor S., Rahman M., Rainville L., Rambani A., Reddy U., Rong J., Saranga Y., Scheffler B., Scheffler J., Stelly D., Triplett B., Deynze A., Vaslin M., Waghmare V., Walford S., Wright R., Zaki E., Zhang T., Dennis E., Mayer K., Peterson D., Rokhsar D., Wang X., and Schmutz J., 2012, Repeated polyploidization of Gossypium genomes and the evolution of spinnable cotton fibres, Nature, 492: 423-427.

https://doi.org/10.1038/nature11798

PMid:23257886

Reddy U., Nimmakayala P., Abburi V., Reddy C., Saminathan T., Percy R., Yu J., Frelichowski J., Udall J., Page J., Zhang D., Shehzad T., and Paterson A., 2017, Genome-wide divergence, haplotype distribution and population demographic histories for Gossypium hirsutum and Gossypium barbadense as revealed by genome-anchored SNPs, Scientific Reports, 7: 41285.

https://doi.org/10.1038/srep41285

PMid:28128280 PMCid:PMC5269598

Rong J., Rong J., Wang X., Wang X., Schulze S., Compton R., Williams-Coplin T., Goff V., Chee P., and Paterson A., 2012, Types, levels and patterns of low-copy DNA sequence divergence, and phylogenetic implications, for Gossypium genome types, Heredity, 108: 500-506.

https://doi.org/10.1038/hdy.2011.111

PMid:22511177 PMCid:PMC3330685

Wang P.P., Dong N., Wang M.J., Sun G.F., Jia Y.H., Geng X.L., Liu M., Wang W.P., Pan Z.E., Yang Q.Y., Li H.G., Wei C.Y., Wang L.R., Zheng H.K., He S.P., Zhang X.L., Wang Q.L., and Du X.M., 2022, Introgression from Gossypium hirsutum is a driver for population divergence and genetic diversity in Gossypium barbadense, The Plant Journal: for Cell and Molecular Biology, 110(3): 764-780.

https://doi.org/10.1111/tpj.15702

PMid:35132720

Wendel J., Schnabel A., and Seelanan T., 1995, An unusual ribosomal DNA sequence from Gossypium gossypioides reveals ancient, cryptic, intergenomicintrogression, Molecular Phylogenetics and Evolution, 43: 298-313.

https://doi.org/10.1006/mpev.1995.1027

PMid:8845966

Wu Y.X., Daud K., Chen L., and Zhu S.J., 2007, Phylogenetic diversity and relationship among Gossypium germplasm using SSRs markers, Plant Systematics and Evolution, 268: 199-208.