Introduction

Sorghum [Sorghum bicolor (L.) Moench] is a member of the Poaceae family, originated in northeastern Africa (Doggett 1988; Kimber 2000). It is widely distributed throughout the world and has various uses, such as in staple foods, animal feed, and as a resource of raw material for bioethanol production (Mathur et al. 2017). Sorghum exhibits wide ecological adaptability, thrives in various climates ranging from arid regions with minimal rainfall to high rainfall and humid areas, as well as temperate regions worldwide (Techale et al. 2022). Moreover, sorghum displays salt tolerance and water-logging resistance (Almodares and Hadi 2009). Compared to prominent cereal crops like wheat, maize, and rice, it exhibits lower input requirements and higher yields (Mathur et al. 2017). With all these advantages, sorghum is considered a “failsafe” crop in agricultural ecosystems (Paterson 2008).

Assessment of genetic diversity in a crop is of high importance in the fields of germplasm management, application, and genotype selection for the purpose of enhancing agricultural development. Genetic diversity can be investigated through molecular markers that can provide rapid and accurate characterization that is unaffected by environmental changes (Jones et al. 2009; Mamo et al. 2023). Various molecular markers, such as Amplified Fragment Length Polymorphism (AFLP) (Gerrano et al. 2014; Medraoui et al. 2024), Random Amplified Polymorphic DNA (RAPD) (Raza et al. 2019; Ruiz-Chutan et al. 2019), Restriction Fragment Length Polymorphism (RFLP) (Cui et al. 1995; Ahnert et al. 1996), Diversity Arrays Technology (DArT) (Mace et al. 2020), Simple Sequence Repeats (SSR) (Zheng et al. 2021; Kedir et al. 2024), and Single Nucleotide Polymorphism (SNP) (Enyew et al. 2022; Yahaya et al. 2023) have been employed in the identification of genetic diversity in sorghum breeding.

Insertion/deletion polymorphisms are a class of markers that arise via many mechanisms, including the insertion of transposable elements, slippage during simple sequence replication, or uneven crossover events (Britten et al. 2003). In comparison to other types of molecular markers, InDel markers have several advantages including abundant presence within the genome, accurate identification capability, reduced demands for high-quality and large quantities of DNA samples for testing, and a simplified experimental process that requires only commonly available equipment (Gao et al. 2012). The number of studies in which InDel markers have been developed and validated has increased significantly because of all these advantages. For example, a large number of InDels was identified in the Mediterranean sesame collection using a ddRAD-Seq protocol developed by Kizil et al. (2020). Out of these, 16 InDels were evaluated across 32 sesame accessions. Liu et al. (2015) developed 48 InDel markers by using conserved sequences derived from functional genes in peanut. These markers were validated in comprehensive panel consisting of 118 different accessions. Ren et al. (2023) identified 195 InDel markers. To further validate, 27 of them were used for 47 double flower cultivars and 22 single flower cultivars of Prunus mume. The process of discovering Indel markers is comparatively sluggish in sorghum when compared to other crops, and there are currently limited studies developing and validating InDel markers in sorghum. Most of InDels were identified in two Sudan grass varieties (Sa and S722) using a RADSeq method (Wang et al. 2022). Out of these, 100 InDels were evaluated across between sorghum Tx623B and Sa and between Tx623B and S722. Choe et al. (2023), two InDel markers with sizes greater than 20 bp were developed for the purpose of discriminating between different cytotypes and validated a total of 1104 plants from six sorghum cultivars to determine variant cytotypes.

Significant advancements have been achieved in sorghum breeding because it has garnered considerable interest in recent years as a potential “star” crop. The Sorghum genome is fully sequenced. This information has contributed to the development of molecular markers such as InDels, which have many advantages, and the assessment of the level of genetic variation which serve as the basis for genetic material selection in plant breeding. The objectives of this study were to: (1) detect InDels using the ddRAD-Seq method and generate PCR-based InDel markers; (2) validate InDel markers in a diverse sorghum panel; and (3) examine genetic diversity and relationships in the sorghum panel.

Materials and methods

Plant materials and DNA extraction

Twenty-seven sorghum accessions (two sweet and one grain sorghum cultivars) were used as genetic materials in this study (Table S1). The seeds of each genotype were sown in pots, and young leaves were collected. The DNAs were extracted using the CTAB method with minor modifications (Doyle and Doyle 1990). The final concentration of DNA was normalized to 100 ng/μL using lambda DNA as a reference.

Library construction and sequencing

A modified version of the ddRAD-seq method (Peterson et al. 2012) was followed for genomic library construction with eleven accessions. Briefly, the genomic DNA was subjected to restriction enzymes; VspI (ATTAAT) and MspI (CCGG) and then Illumina paired-end adaptors (P1 and P2, P1 adapter's 3' end was modified to match the overhanging VspI restriction cut site) were ligated to the fragmented DNA using T4 DNA ligase buffer (Peterson et al. 2012). PCR was used to amplify with genotype-specific indexed PCR primers. The ligated DNA products were visualized on 2% agarose gel and selected based on the average size of 400–500 bp DNA fragments. The resulting ddRAD-seq genomic library was sequenced on an Illumina Hi-Seq platform via 150 bp paired-end sequencing.

InDel calling

InDel calling was performed in the following steps:

  1. 1.

    The raw reads were demultiplexed with Je V1.2 (Girardot et al. 2016),

  2. 2.

    Trimming of adaptors and filtering was conducted with fastp (v0.20.1) (Chen et al. 2018) (Phred quality score less than 15 out of 40),

  3. 3.

    The filtered reads were aligned with sorghum reference genome (McCormick et al. 2018) using Bowtie2 (v2.3.4.3) (Langmead and Salzberg 2012),

  4. 4.

    The created BAM files analyzed, and variants were called with freebayes (v13.1) (Garrison and Marth 2012) (simple diploid calling and filtering, and coverage values of 20x),

  5. 5.

    VCF files were filtered with (SNPs were discarded) VCFfilter (Galaxy Version 1.0.0),

  6. 6.

    All.vcf files were combined with (individual.vcf files included insertions and deletions; with VCF genotypes (Galaxy Version 1.0.0).

  7. 7.

    The merged variant file underwent processing in Microsoft Excel to remove duplicate areas and arrange the InDels based on their sizes.

  8. 8.

    The InDel regions with a minimum length of 10 base pairs, were evaluated using the Integrated Genome Browser V9.1.4 (IGB) (Freese et al. 2016). This analysis was conducted for each BAM file in conjunction with the sorghum reference genome.

Primer design and PCR analysis

To generate genome-wide InDel markers, flanking sequences of selected insertion-deletion regions were used. Primers were designed with Primer3Plus (Untergasser et al. 2007) with the following parameters: predicted PCR products ranged from 100 to 900 bp; the length of primers was limited to 18 bp; the annealing temperature was restricted to 50–58 °C. The primer pairs were subsequently checked with Integrated Genome Browser V9.1.4 (Freese and Norris 2016), for possible sequence duplications in the genome. All markers were given a name with the format SG-D(I)-X-XXX, where "SG" defines sorghum, "D" and "I" indicate deletion and insertion, "X" is the chromosome number, and "XXX" denotes the beginning of the chromosomal position. The InDel annotation was derived from the reference genome sequence of sorghum using Generic Feature Format 3 (Sorghum_bicolor_NCBIv3.54.gff3).

The PCR analyses were performed in 15 μl of reaction mixture containing 1 μl DNAs (40 ng), 1.5 μl 10 × PCR buffer, 2.5 mM MgCl2, 1.5 μl dNTP mix (4 mM), 1 μl each of forward and reverse primers (10 μM), 0.2 μl of Taq DNA polymerase (5 U/μl), and amount of Milli-Q water. The thermal cycles include 94 °C for 5 min, 30 cycles of 94 °C for 59 s, 50–58 °C for 59 s and 72 °C for 1 min, with an extension 72 °C for 10 min. The PCR products were analyzed by 3% agarose gel electrophoresis and visualized by UV light.

Validation of developed InDel markers

A total of 16 accessions with diverse genetic backgrounds were screened with the developed 14 InDel markers. The predicted PCR bands were visualized with a Fragment AnalyzerTM (Advanced Analytical Technologies GmbH, Heidelberg, Germany) with the use of DNF-900-K0500 reagent kit for the qualitative analysis of DNA fragments. The solutions, buffers, and gels were used in accordance with the instructions provided by the manufacturers. The bands were visualized in virtual gel which show size of bands ranging from 1 base pair (bp) as the lowest and 500 bp as the highest. Additionally, calibration was performed using a DNA ladder within the 1 to 500 bp range. The obtained data was evaluated using PROSize 2.0 (Advanced Analytical Technologies, AMES, IA, USA).

Genetic diversity analysis

Population genetic parameters, including number of alleles (Na), effective number of alleles (Ne), Shannon diversity index (I), expected heterozygosity (He), and principal coordinate analysis (PCoA) were calculated using GenAlex V6.5 (Peakall et al. 2012). The polymorphism measurement was performed using the Excel Microsatellite Toolkit (Park 2002).

Results

After quality filtering, a total of 33.7 M reads was obtained by paired-end sequencing. The mean number of reads was 3.06 M and the guanine-cytosine (GC) content was 40.32% among the sequenced genotypes. The highest and lowest reads were acquired by BSS47 genotype and Ogretmenoglu variety, respectively. The mean of the ddRAD reads mapped to the Sorghum bicolor reference genome was 81.89%.

We identified 19,226 InDel sites following to the bioinformatic pipeline. The deletions were composed of 65.7% of the total InDel positions, and their sizes varied from 1 to 14 bp, and 90.8% of all deletions had a size length of 1 to 2 bp. Among the insertions, 95.4% were less than 5 bp length, 4.5% were between 5 and 10 bp, and 0.3% were more than 10 bp (Table 1).

Table 1 The number of InDels determined within ddRADSeq analysis of eight sorghum accessions, two sweet sorghum and one grain sorghum cultivars
Full size table

A total of 4814 InDel regions were identified, each with a size of ≥ 2 bp, distributed throughout the genome. The number of markers per chromosome ranged from 347 to 730 (Table 2).The identified deletion markers were 2903which were higher than the number of insertion markers which was 1911. Furthermore, variations in the distribution of insertions and deletions were observed across different chromosomes. Chromosome 1 had the highest number of deletions with 483, while chromosome 8 had the lowest number of insertions with 150. The frequency of InDels (size ≥ 2 bp) varied among the chromosomes, ranging from 5.53 to 9.02 InDels/Mb. The highest frequency of deletions and insertions was observed on chromosomes 1 and 2, while chromosome 8 showed the lowest frequency (Table 2).

Table 2 Distribution of insertions-deletions (size ≥ 2 bp) in genome
Full size table

A total of 80 InDel sites (≥ 10 bp) was detected in the sorghum genome (Table 3). Chromosome 10 exhibited the highest number of deletions (9), while chromosome 2 displayed the greatest number of insertions (10). There was no deletion and insertion of ≥ 10 bp in chromosomes 1 and 3, respectively (Table 3). Chromosomes 8 and 9 each exhibited only one deletion with a length of ≥ 10 bp. The chromosome 8 exhibited the longest insertion, spanning 14 bp, at a physical location of 54,175,764 (Table S2). On the other hand, the longest deletion, spanning 14 bpeach, was shown in chromosome 4 (physical position: 10,180,777), chromosome 5 (physical position: 10,427,959), and chromosome 6 (physical position: 47,562,364) (Table S3). The Integrated Genome Browser (IGB) also visualized the genomic positions of selected insertions and deletions (Figs. S1 and S2).

Table 3 Distribution of insertion-deletions with a length of ≥ 10 bp in the genome
Full size table

The primer pairs were designed with the use of 47 insertion-deletions that were identified across all the 10 chromosomes that have a length of 10 bp and/or greater. The highest number of markers was designed in chromosome 2 (9 markers), followed by chromosome 10, with 8 markers in total, including 7 deletion and 1 insertion marker. The primers used in this study successfully generated clean amplicons, as shown in Fig. S3. The analysis of insertion-deletion genomic positions revealed that the majority of those (74.6%) was located in intergenic regions. A smaller proportion was found in messenger RNA regions (mRNA) (12.7%), introns (6.4%), exons (4.2%), and coding sequences (CDS) (2.1%) (Table S4).

We tested 14 InDel markers (lengths ≥ 10 bp) with genetic diversity analysis using 16 sorghum accessions that originated from nine countries. The PCR products varied between 100 and 600 bp, which were visualized on a Fragment Analyzer® (Fig. S4). The InDel markers exhibited predicted polymorphisms among the accessions and were submitted in Table 4. The expected heterozygosity was determined to range from 0.000 to 0.135. The loci SG-D-2-581 and SG-I-6-563 exhibited the greatest expected heterozygosity value. The average expected heterozygosity across all loci was determined to be 0.072. The average Shannon diversity index (I) was determined to be 0.101. The range of polymorphic information content (PIC) values for these 14 markers varied from 0.110 to 0.380, with an average value of 0.282. The results of Principal Coordinate Analysis (PCoA) revealed that the first and second coordinates accounted for 26.50% and 17.04% of the overall variance, respectively. The sorghum accessions were further categorized based on the inclusion of accessions originating from various continents; they were divided into three groups in the PCoA graphic. The first group (I) included seven genotypes (BSS206, BSS224, BSS241, BSS244, BSS283, and BSS293) from three continents. The second (II) and third (III) groups contained four (BSS249, BSS269, BSS270, and BSS295) from four continents and five (BSS203, BSS210, BSS221, BSS232, and BSS292) from one continent, respectively (Fig. 1).

Table 4 Summary of genetic diversity statistics for 16 sorghum accessions
Full size table
Fig. 1
figure 1

Principal coordinate analysis (PCoA) of the 16 sorghum accessions genotyped with 14 InDel markers

Full size image

Discussion

With the advent of next-generation sequencing (NGS) technologies, different types of genome reduction methods were developed, which has provided effective opportunities for generating data for marker development. ddRAD-Seq is one of the reduced-representation sequencing approaches to identify variations in the studied genome. The method provided lots of data to identify the genetic diversity and population structure of the studied sorghum panel. In this study, we identified a total of 19,226 InDel loci based on ddRAD-Seq. The distribution of these loci across chromosomes confirmed the suitability of this approach for the development of genome-wide molecular markers. This genome reduction method was also used to develop markers successfully in sesame (Kizil et al. 2020) and lettuce (Seki 2022). However, this study is the first to report the suitability of the ddRAD-Seq approach in the development of InDel markers in sorghum.

Table 1 illustrated the identification of 14 InDel types. InDels that were 10 bp or longer made up 0.4% of all the identified InDels. Generally, as the size of an InDel increased, InDel numbers decreased in the sorghum genome. Furthermore, single-nucleotide InDels were predominant (Table 2), which is consistent with the previous findings in sesame (Kizil et al. 2020), ginkgo (Wang et al. 2023), and soybean (Song et al. 2015). On the other hand, maize (Batley et al. 2003) exhibited a higher prevalence of bi-nucleotide InDels. Our analysis revealed that chromosome 1 had the highest number of InDels with a minimum length of 2 bp. This finding aligns with prior research on sorghum, which indicated that chromosome 1 exhibited the highest abundance of InDels (Li et al. 2016) and SSRs (Shehzad et al. 2020). These findings may also be attributed to the increased likelihood of survival among plants with shorter InDel lengths. Both insertion and deletion events may have negative impacts on the genome, resulting in the loss of some genes and ultimately causing harm to plants. When the length of InDel is shorter, the resulting harm to the plant may be less severe, hence increasing the likelihood of its survival. Thus, the probability of preserving this variation increases (Liu et al. 2019a). Moreover, no deletions and insertions of 10 base pairs or longer were observed on chromosomes 1 and 3, respectively. This could be a drawback of ddRAD-Seq, resulting in large gaps in genome coverage following the sequencing of a prepared genomic library (Shirasawa et al. 2016).

The present study revealed that a majority of the InDel events occurred inside non-coding regions (Table S4), similarly to the findings of Zheng et al. (2011), who developed InDel markers from sweet and grain. The proportions of InDel primers developed from regions of mRNA, introns, exons, and CDS were 12.7%, 6.4%, 4.2%, and 2.1%, respectively (Table S4). InDels, which occur inside functionally significant areas of genes, particularly coding sequence (CDS) regions, have the potential to impact gene function by causing frameshifts and altering the structure of proteins (Zhang et al. 2016). The results of the study indicated that there was a limited presence of InDels inside the CDS (14 bp deletion, position; 47,562,364) area. This result may be attributed to the relatively small proportion of the CDS region represented in the analysis as well as its higher degree of conservation compared to other genomic regions (Liu et al. 2019b). Moreover, InDel within CDS regions has a propensity to have a substantial influence on the structure and function of proteins (Liu et al. 2019b) and can be considered a valuable resource for developing phylogenetic and/or functional markers (Ramakrishna et al. 2018). For example, researchers previously published the identification of the InDel-1891 marker, which has the capability to distinguish restorer lines that possess the Rf1 gene in cotton (Wu et al. 2017).

InDels are the second most prominent kind of structural variation after SNPs uncovered in the plant genome (Yang et al. 2014). These markers have robustness and efficiency comparable to SSRs (Shi et al. 2023) and may be chosen based on their expected fragment size and then confirmed in genetic populations or germplasm via the use of a simple and cost-effective agarose gel electrophoresis technique. Previously, InDel markers have been used for the purpose of analyzing genetic diversity in many crops, including sesame (Kizil et al. 2020), chickpea (Jain et al. 2019), sweet potato (Zhao et al. 2022) and rice (Sahu et al. 2017). However, there are a limited number of studies in which InDel markers have been evaluated for genetic diversity in sorghum (Zheng et al. 2011; Boatwright et al. 2022). In this study, we developed lots of markers that can easily show the polymorphism in agarose gels, and 14 InDel markers were selected to identify genetic variation in 16 sorghum accessions with diverse origins. The polymorphic information content (PIC), which offers an estimation of the information content associated with a marker, is one of the most common measures of the polymorphism of markers (Shete et al. 2000). The analysis of genetic diversity found the highest PIC value was 0.380. The average PIC value of 0.282 for a set of 14 markers is higher than the value reported for the SNP markers used by Yahaya et al. (2023) and Enyew et al. (2022) in sorghum. The results of this research indicated that the InDel markers had a high ability to reveal genetic diversity in the collection. The observed mean values for He (0.07) and I (0.01) were indicative of low genetic variation within the sorghum accessions. In general, previous study on sorghum have shown relatively low genetic variation among landrace accessions (Motlhaodi et al. 2017). The observed result may be attributed to a mix of its inbreeding nature and the rigorous selection criteria used by farmers. In this study, the principal coordinate analysis (PCoA), which explained 26.50% of the total variation in its first two principal axes, provided three class in relation to geographical origin. Accessions were mostly based on the geographical origins in the first and third groups, with some exceptions. This condition may arise from the hybridization of different varieties across continents, facilitating the exchange of cross-gene flows. Moreover, these exceptions indicate that there is a high probability of sorghum landrace genotypes being shared throughout different areas by farmers, perhaps via various means of exchange. This observation is consistent with the idea of seed mixing, exchange, and trade among small-scale farmers, which underscores the dynamic character of genetic connections and seed mobility within agricultural communities.

Conclusion

Next generation sequencing of sorghum accessions with a kind of genome reduction method of ddRAD-seq made possible to develop a large number of InDel markers in sorghum. The findings of this study suggested that the use of this technology is an efficient for the development of genome-wide markers. Among the 80 InDel sites with a minimum length of 10 base pairs, 47 markers that could be resolved using agarose gel electrophoresis were effectively amplified. Due to its ease of detection and cost-effectiveness via the use of readily accessible PCR equipment, these markers may be significant value in genetic analysis and breeding programs pertaining to sorghum.