G. hirsutum (Linnaeus, 1753) accession TM-1 was used for cytological studies. BACs used for FISH mapping were identified by screening two genomic BAC libraries derived from G. herbaceum (Linnaeus, 1753) var. africenum (Gao et al. 2013) and G. barbadense (Linnaeus, 1753) Pima 90-53 (kindly provided by Prof. Zhiying Ma of Hebei Agricultural University). The chromosome-specific BAC clones for G. hirsutum A_h01/D_h01 were kindly provided by Prof. Tianzhen Zhang of Nanjing Agricultural University, The simple sequence repeat (SSR) markers used for BAC screening were selected from a whole genome marker map (WGMM) (Wang et al. 2013) and a genetic map (Yu et al. 2011).

The screening was performed using bacteria liquid-PCR according to the protocol previously described (Cheng et al. 2012).

Chromosome preparation and FISH were conducted according to the previous protocols (Gan et al. 2011). In order to reduce the interference from the background signals, heat-shock-interrupted (1.5 mL Eppendorf tube filled with 100 μl genome DNA was placed in sterilization pot with 105°C for 8 min) cotton genome DNA fragments with size from 200 bp to 800 bp were used as blocking DNA. BAC-DNA used to label probes was isolated using Plasmid Miniprep Kit (Biomiga) according to the handbook. Biotin- and digoxigenin-labeled probes were detected using rhodamine-conjugated anti-digoxigenin and fluorescein-conjugated avidin (Roche Diagnostics, USA), respectively. Chromosomes were counter-stained with 4, 6-diamidino-2-phenylindole (DAPI, Sigma, USA) and antifade (Vector, USA) under a cover-slip.

Slides were examined under a Zeiss Imager M1 microscope. Images were captured and merged using MetaSystems isis software with a CCD camera (MetaSystems CoolCube 1) attached to a Zeiss Imager M1 microscope. To determine physical positions of signals, only chromosomes without apparent morphological distortion were introduced and their physical positions of signals were measured using MetaSystems isis. Final image adjustments were performed using Adobe Photoshop CS3 software.

The RMP unit was used as standardized map unit for comparative analysis between different types of maps. The RMP values for the SSR linkage map were the percentage from the genetic location (cM) of each locus along the total length (cM) of the corresponding linkage group. The RMP values of the cytogenetic map were the percentage of the distance (μm) from the FISH signal site to the end of the short arm showed relative to the total length of the chromosome (μm) (Sun et al. 2013). In order to determine the genomic locations (bp) of each BAC clones, the primer sequences of BACs-corresponding SSR markers were obtained from the database Cotton Marker Database (http://www.cottonmarker.org/), then according to Electronic PCR command line tools (Version 2.3.12), e-PCR was run against the G. hirsutum (AD₁) genome NAU-NBI Assembly (https://www.cottongen.org/organism/Gossypium/hirsutum) according to the default parameters. The RMP values for the G. hirsutum draft genome assembly map were calculated from the genomic location (bp) of each locus along the physical length of chromosomes A_h01 and D_h01. These RMP values were used to produce the comparative map alignments.

To construct the cytogenetic maps of chromosomes A_h01 and D_h01 of G. hirsutum, an initial set of 47 SSR markers shared by both chromosomes of A_h01 and D_h01 from a whole genome marker map (WGMM) (Wang et al. 2013, Rong et al. 2004) and a genetic map (Yu et al. 2011) were used to screen two BAC libraries of G. herbaceum var. africenum and G. barbadense Pima 90-53. Based on the WGMM, the SSR markers were distributed along the linkage group of chr.15 (D_h01) from 0.6 cM (CIR009) to 176.3 cM (CIR110) (Table 1). In total, 84 positive BAC clones were identified based on the result of BAC libraries screening (Table 2). Due to abundance of repetitive sequence in cotton genome, by dual-color FISH with the chromosome-specific BAC clones 52D06 (A1) and 48F11 (D1) as controls, only 12 BAC clones were selected for FISH mapping which produced little or no background signal when hybridized to G. hirsutum chromosomes with the aid of blocking DNA.

By dual-color FISH on mitotic chromosomes, the order of the two BACs was determined along the chromosomes based on the genetic positions of their corresponding SSR markers. Results showed, among the 12 positive BAC clones, 11 BAC clones were homoeologous-specific BACs because they generated signals on both chromosomes of A_h01 and D_h01, indicating sequence homology between these BACs retained in A_h01 and D_h01 (Fig. 1a-k). One BAC clone 378J07, derived from SSR HAU076, only had one pair of FISH signals on chromosome A_h01, which had collinearity with the chromosome A_h01-specific BAC clone A1 (52D06) (Fig. 1l). Based on these results, the relative position of all probes can be preliminarily distinguished along the mitotic metaphase chromosomes.

Figure 1.

The order of two BACs on metaphase chromosome of G. hirsutum (AD₁) TM-1 using Dual-color-FISH. a 305A19(green)/348I20(red) b 305A19 (green)/64M24(red) c 144E04 (red)/64M24(green) d 64M24 (red)/400N03(green) e 305A19 (red)/378J07(green) f 118G12(red)/164I24(green) g 423C18 (red)/400L15(green) h 85P13 (green)/400L15(red) i 85P13 (red)/421E24(green) j 85P13 (red)/164I24(green) k D1 (red)/118G12(green) l 378J07 (green)/ A1 (red). Bar = 5 µm.

The genetic distances of SSR markers associated with the corresponding BACs were also converted into the relative positions in the corresponding linkage map (Fig. 2a). In order to confirm the physical position of each clone, FISH signal of each BAC clone was measured in 5-8 cells with clear chromosome spreads and the RMP of FISH signals were computed (Table 3). Based on the data, the cytogenetic maps of the homoeologous chromosomes D_h01 and A_h01 were constructed (Fig. 2b, c). The order of individual BACs along the chromosome was generally collinear with the order of the corresponding SSR markers along the linkage map, except for a few closely linked loci, 144E04 (NAU2474) and 348I20 (NAU3253), 118G12 (NAU4891) and 400N03 (BNL2921), which displayed changes in the order between the genetic markers and BAC locations (Fig. 2a, b). Moreover, the BACs showed better concordance in the orders and positions between the two cytogenetic maps of the homoeologous chromosomes A_h01 and D_h01, except for 400N03 (BNL2921) (Fig. 2b, c), which suggests a rearrangement between the A_h01 and D_h01 homoeologous chromosomes in the process of evolution. A significant difference between the two types of maps was viewed, that is, the markers flanking the middle region were separated by short genetic distance but long physical distance. For example, the genetic distance between markers NAU3433 and BNL2921 is 11.2% of total genetic distance of chromosome 15 (D_h01), but the physical distances between these two markers is 59.4% of the total length of the chromosome D_h01 (Fig. 2a, b).

Figure 2.

Comparison of positions of BACs in cytogenetic maps of G. hirsutum A_h01/D_h01 with genetic positions of SSR markers a Positions of SSR markers based on WGMM; c, b Cytogenetic maps of G. hirsutum A_h01/D_h01.

To compare our cytogenetic maps directly to the draft genome assembly map (Zhang et al. 2015), the corresponding SSR primers of the BAC clones were mapped to the draft genome sequence by e-PCR, and the relative positions of the SSRs were calculated according to the e-PCR results (Table 3). Based on the above data, we integrated the cytogenetic maps with the genome sequence maps of the homoeologous chromosomes A_h01 and D_h01 to compare their distributions (Fig. 3). The alignments allowed a global view of the relations between the chromosomal positions and physical positions in draft genome map of the BAC clones. The number of BACs mapped on each pseudo-chromosome in the draft genome assembly map was significantly less than that on the corresponding cytogenetic maps (six to twelve on D_h01, nine to twelve on A_h01) (Fig. 3). Of the eleven homoeologous-chromosomes-shared BACs based on cytogenetic maps, four BACs’ corresponding SSR markers (NAU3433, NAU3135, NAU2474 and NAU4044) were simultaneously mapped on the two corresponding chromosomes in G. hirsutum draft genome assembly. The others were only mapped on one of the chromosome A_h01 or D_h01 respectively. NAU2474 was mapped on the chromosome D_h01 and scaffold183_A01 of the draft genome assembly by e-PCR. Its corresponding BAC clone 144E04 was FISH mapped on chromosome A_h01 (RMP 9.01%) and D_h01 (RMP 4.33%) in cytogenetic maps. NAU4044 was mapped on the chromosome A_h01 and scaffold3710_D01 of the draft genome assembly by e-PCR. Its corresponding BAC clone 400L15 was FISH mapped on chromosome A_h01 (RMP 96.01%) and D_h01 (RMP 90.33%) in cytogenetic maps. Based on these comparison results, the locations of the two scaffolds in the draft genome assembly were determined approximately. That is, scaffold183_A01 (size 55529 bp) located between the SSR markers HAU076 and NAU3433 on the chromosome A_h01, i.e., the relative position between 3.4% and 9.6% (sequence loci from 90268610 bp to 96488204 bp) (shown by arrow Fig. 3d). Scaffold3710_D01 (size 191022 bp) locates near the end of the chromosome D_h01, i.e., the outer of the relative position 84.7% (sequence loci from 6145600 bp to 9387374 bp) (shown by arrow Fig. 3a).

Figure 3.

Integrated cytogenetic /genome assembly maps of G. hirsutum A_h01/D_h01. a Relative map position of BACs mapping to D_h01 of the AD₁-NBI draft genome b Cytogenetic map of G. hirsutum Chromosome D_h01 based on 12 BAC clones c Cytogenetic map of G. hirsutum A_h01 based on 12 BAC clones d Relative map position of BACs mapping to A_h01 of the AD₁-NBI draft genome. Arrow-head in a and d represent the locations of scaffold3710_D01 and scaffold183_A01 in the draft genome (AD₁-NBI) respectively.

In cotton, more than 30 genetic maps have been published, including several integrated maps with higher marker density (Yu et al. 2010, Yu et al. 2011, Blenda et al. 2012), and a whole-genome marker map (WGMM) by integrating publicly available sequence tagged DNA markers with the cotton D-genome sequence (Wang et al. 2013). Undeniably, they are a foundational tool and resources for marker-assisted selection and genomic studies. But the linkage maps provide little information about physical locations, distributions, distances, and sometimes orientations of genetic markers. Cytogenetic maps encompassing the information from both genetic maps and cytological maps, can relate the markers mapped across linkage groups to cytological position on chromosomes. Using a set of marker-anchored BACs, we developed the cytogenetic maps of homoeologous chromosomes Ah01 and Dh01 in G. hirsutum. The comparative map alignments revealed a significant disproportion between genetic and physical distances in the pericentromeric region, such as, the distance between markers NAU3433 and BNL2921 with 11.2 RMP(Fig. 2a) but on the cytogenetic map with 59.4 RMP (Fig. 2b). The reduction of recombination around the chromosome centromere is a common feature and the region of recombination suppression correlates directly with sizes of centromeric heterochromatic regions (Sun et al. 2013). So this implies larger region of suppressed recombination was detected in the pericentromeric region of chromosome D_h01. Moreover, the orders of most genetic markers are collinear with corresponding BAC locations although several closely linked loci in D_h01 display inconsistent orders or locations compared with those in BAC FISH maps.

In total, the integrated genetic and cytogenetic maps can serve as a template to facilitate sequence assembly, because the maps provided information on the distribution of genetic markers across chromosomes and the linkage gaps derived from recombination suppression.

As a typical allotetraploid, which contains two sub-genomes originating from related ancestor species with different genome sizes, G. hirsutum has been studied on its homoeologous chromosomes for a long time. Results revealed that fragment additivity (Liu et al. 2001), the independence of evolution of duplicated genes (Cronn et al. 1999), conservation in gene content, order, and spacing (Grover et al. 2004, 2007) between the homoeologous chromosomes, as well as the potential mechanisms for genome-size variation in the homoeologous chromosomes (Wang et al. 2010). Here, we constructed the cytogenetic maps of homoeologous chromosomes A_h01 and D_h01 using shared-markers-anchored BACs. By comparison analysis of BACs’ positions, consistent orders of FISH signals were viewed in both homoeologous chromosomes, except for one BAC clone 400N03, which showed obvious location discrepancy in the homoeologous chromosomes (RMP 62% in A_h01 and 84.7% in D_h01). The discrepancy may be caused by a chromosomal rearrangement in this region during a certain period of polyploidization. In addition, better collinearity of ten of eleven shared BACs between the homoeologous chromosomes suggests that there remains a generally high level of sequence conservation between homoeologous chromosomes A_h01 and D_h01, though polyploidization occurred about 2 MYA (Cronn et al. 2002, Seelanan et al. 1997, Wendel 1989).

The e-PCR can be used to search for sub-sequences that closely match the primers of SSRs, which can help to identify the genome positions of SSRs within the reference genome sequence (McCouch et al. 2002, Li et al. 2015). In this study, we identified the genome positions of thirteen SSRs using e-PCR. Results showed the length and position of the target sequence for each pair of primers against the reference genome sequence were consistent with the initial selection, which ensured the accuracy of the next relative position calculation and comparative analysis.

Mis-assemblies are common when draft genome sequences have been generated by de novo assembly of sequences obtained with NGS technologies (Meader et al. 2010, Alkan et al. 2011). Since the assembly of G. hirsutum was done using the SOAPdenovo software, the final assembly comprised 265,279 contigs and 40,407 scaffolds (Zhang et al. 2015), so mis-assembled scaffolds may exist in the draft genome. On the other hand, there are a generally high level of sequence conservation between homoeologous genomic regions in allotetraploid species including cotton and wheat (Zhao et al. 2012, Brenchley et al. 2012), it is difficult to annotate and assemble whole-genome sequences. Since the cytogenetic map can reflect the true position of the DNA sequence in the chromosome, so it has some significance for verification and correction of the genome assembly. In the process of genome sequencing and sequence assembly, the cytogenetic map plays a role in filling the sequencing gaps, correcting assembly errors, evaluating the quality of assembly, achieving more scaffolds and contigs chromosomal localization and orientation. Wang et al. mapped 32 BAC clones to some of the homologous chromosomes 12A and 12D of upland cotton by FISH, and constructed the high resolution cytogenetic map of the two chromosomes (Wang et al. 2010). Through the integration of genetic loci and physical sites, considerable variations in the composition, structure and size of the two homoeologous chromosomes were viewed, which play an important role in the sequencing and sequence assembly of G. hirsutum (Wang et al. 2010; Zhang et al. 2015). By comparison of the distributions of fosmid clones on the cucumber draft genome assembly map and cytogenetic map, the accuracy and coverage of the draft genome assembly map were evaluated (Sun et al. 2013).

Here, we constructed the cytogenetic maps of homoeologous chromosomes A_h01 and D_h01 using shared-BACs. By integration of cytogenetic maps and the cytogenetic and genome assembly maps, we identified the positions of two scaffolds in chromosome (Fig. 3a, d). Among the eleven shared-BACs in the cytogenetic maps of chromosomes A_h01 and D_h01, only four (accounting for 36.36%) had hits both in two corresponding pseudo-chromosome in the draft genome assembly map, the others were only mapped on one of the chromosome A_h01 or D_h01 respectively. It may be that some homologous sequences were removed as repeats, and only partial sequences information with homology were assembled on one of the two homoeologous chromosomes during the assembly process.