Survey sequencing and in-silico development and validation of genomic SSR markers in Indian dill seed

2.1 DNA isolation and DNA sequencing

To isolate microbial genome free DNA of dill seed, the cultivar GAVD-1 was gown in the controlled condition in seed germinator (Percival). For high-quality DNA isolation, leaves from aseptically grown 15 days old seedlings were collected. The CTAB protocol of Doyle and Doyle (1990) was used for genomic DNA isolation. Isolated was assessed on 0.8% agarose gel and spectrometer to confirm the intactness and quantity of extracted genomic DNA.

PCR-Free TruSeq (LT) DNA sample Prep kit (FC-121–3001) was used to generate the library following the protocol recommended by Illumina. Paired-end sequencing was performed in the departmental laboratory on MiSeq platform (Illumina) with MiSeq Reagent Kit version 2 (2 × 250 bp). The details on up- and down-stream procedure for DNA sequencing is given in Bharti et al. (2018).

2.2 Bioinformatic analysis for SSR marker development

During downstream processing of raw sequences, quality checking was carried out using FastQC tool (Andrews, 2010) for confirmation of high-quality reads. For removing adaptor sequences and trimming of low-quality bases Trimmomatic (Bolger et al., 2014) was used. During raw reads processing, reads were filtered with the custom setting (Phred 33, Sliding window:4:15, Leading: 10, Trailing: 10, Minlen: 35) to remove adaptor sequence or low-quality reads. After trimming, FastQC analysis was again performed to check the high quality of trimmed reads. Subsequently, high-quality clean reads were utilized for de-novo assembly using SPAdes-3.11.1 (Bankevich et al., 2012), Megahit -1 (Li et al., 2015) and SOAPdenovo 2.04 (Luo et al., 2012) software. SPAdes was performed with the read error correction program. QUAST was used to evaluate the results of each assembly (Gurevich et al., 2013). Eventually, the best assembly was used as an input file for MISA to detect microsatellite using default parameters and primer3 was used to generate SSR markers.

2.3 Marker validation and cross-amplification

For validation of primers, a total of 12 SSR primers were selected randomly. The DNA of dill seed cultivar GAVD 1 was used for primer amplification. The DNA amplification was achieved through PCR in a final volume of 15 μl with 1× PCR buffer, 1 μlMgCl2, 0.5 μl dNTPs, 1 μM of primers (10 pmol), 0.3 μl of Taq DNA polymerase (5 U/μl), and 20 ng of template DNA. PCR reactions were completed with initial denaturation at 94 °C for 5 min, followed by 35 cycles of 94 °C for 30 s, primer-specific annealing temperature (Ta) for 40 s, and 72 °C for 60 s with a final extension of 10 min at 72 °C. The amplified products were separated in agarose gel (1.5%) and photographed in a gel documentation system (G-box, Genei). For cross-amplification of SSR primers, DNA six seed spices (Ajwain, Celery, Coriander, Cumin, Fennel and Parsley) belonging to the Apiaceae family was extracted followed by PCR as described above. The amplification products were scored in base pair on a visual basis. For evaluation of SSR marker transferability, the microsatellite products were classified into two classes based on the amplicon intensity and scoring ease: (++) strong amplicon and easy to score and (+) weak amplicon and hard to score. The amplicons were declared as non-specific if the product is of dissimilar size (within +100 bp) to that of dill seed.

3.1 Filtering of reads and de novo genome assembly

Survey sequencing of the library generated 2119.51 Mbp of raw data containing 7,102,148 paired-end reads. The average read length was ∼298 bp. After cleaning and filtering the raw reads, 6,516,297 reads were obtained from high-quality 1541.62 Mbp. The mean read length of clean data was ∼207 bp which is similar to Shiwani (2017) where it was 203 bp. The GC content of filtered reads was 34%.

Evaluation of all three assemblies through QUAST suggested that assembly generated by SPADES was best. Using high-quality reads, SPADES generated 2,25,956 contigs (183.54 million bases). A set of 156,521 contigs were of ≥500 bp long. The maximum and minimum contig length was 92,006 and 128 bp, respectively. The N50 was 604 and 832 with SOAP denovo and Megahit, respectively. However, the N50 value of this study was 1020 bp in case of SPADES assembly. Compared to reported of Shiwani (2017), where N50 in dill seed was 661 bp, N50 was 1.5X higher. The GC content (32.46%) in the current study was closer to carrot (34.8%) and coriander (34.65%) as reported by Iorizzo et al. (2011) and Tulsani (2013).

3.2 Characteristics of genomic SSRs

A total of 37,863 (16.75%) contigs, out of 2,25,956 were containing microsatellite motifs. Earlier, in cumin, SSRs motif in survey sequences was 12% of the total contings (Bharti et al., 2018). A total of 48,951 repeat motifs were detected suggesting one SSR/3.75 kb considering 183.54 million bases. Moreover, 8185 contigs were found to be containing more than 1 SSR, and 5114 contigs were identified having SSRs present in compound formation (Table 1). Previously, in fennel, 954 (18%) transcripts exhibited more than one microsatellite region (Palumbo et al., 2018).

Among detected repeat motifs, the 48.89% (23936) was mononucleotides. Compared to cumin (15%) and celery (21.66%) genome, the mononucleotide repeats was very high (Li et al., 2014a; Bharti et al., 2018). The frequency of A/T SSRs was more in mononucleotide repeats, almost 27x more than C/G SSRs which is in congruence with Li et al. (2014a). After mononucleotide, the majority (49.02%) of SSRs were detected in the di- and tri-nucleotide categories (37.34% and 11.67%, respectively). Penta- and hexa-mers were least common motifs (0.24% and 0.11%, respectively) (Table 2). This is in agreement with the previous report in various seed spices crops where di- and tri-mers were frequent while penta- and hexa-mers are exceptional (Li et al., 2014a; Bharti et al., 2018; Palumbo et al., 2018). Similar to Li et al. (2014a) in celery, the repeat-motif between 5 and 8 were dominant, accounting for 56% of all the SSRs identified. AT/AT was the greatly frequent (64.91%) dimer repeat whereas AC/GT and AG/CT were present in equal frequency (17% each) (Table 3). The similar pattern was also recorded in cumin (Bharti et al., 2018). Many researchers have also recorded richness of AT/AT repeat motif in the genome of different crops (Kalia et al., 2011; Xu et al., 2013). The less existence of CG/CG motif also demonstrated depicted in different Apiaceae crops high (Li et al., 2014a; Bharti et al., 2018).

An overabundance of both di- and tri-nucleotide repeats in cumin, celery and parsley -members of Apiaceae- has been recorded (Li et al., 2014a,b; Bharti et al., 2018). The overabundance of di- and tri-mer repeats disagree with SSRs in Triticum aestivum and Saccharum spp, wherein tetranucleotide repeats were dominant (Cordeiro et al., 2001; Gupta et al., 2003). In the current study, dinucleotide repeats were higher frequent that trinucleotide repeats. This is in agreement with findings in citrus (Biswas et al. 2014) and sesame (Wei et al., 2014).

3.3 Validation of the SSR markers and their transferability

A small set of 12 SSR primers were used for amplification to validate the primers. Of 12 primers, 10 (83%) primers could successfully amplify in dill seed. Representative amplification profiles of few primers are shown in Fig. 1. High amplification rate suggested that assembly was perfect and primers have a proper binding site in the genome. The amplification rate has also been reported by Bharti et al. (2018) in cumin for genomic SSR markers. Amplicon size of validated primers ranged from 120 to 260 bp. A total of 11 alleles were amplified by ten primers in dill seed (Table 4).

Close

Fig. 1

Amplification profile of genomic SSR markers in members of family Apiaceae; M = 50 bp ladder, 1 = Dill seed; 2 = Ajwain; 3 = Celery; 4 = Coriander; 5 = Cumin; 6 = Fennel; 7 = Parsley.

Export to PPT

Most of the member of family Apiaceae are deficient in genomic resources and molecular markers (Kumar et al., 2014). Most of the (EST)SSR markers are produced through transcriptomic study in fennel, celery, parsley etc. (Li et al., 2014a, b; Palumbo et al., 2018). EST-SSRs are superior to genomic SSRs due to a segment of transcripts. However, EST-SSRs are less polymorphic due to the selection against the variation in the conserved regions of the EST-SSRs (Zhang et al., 2014). This suggests that there is a need to develop genomic SSR markers in the above crops as well as other members of the carrot family. Though SSR development with NGS is cost-effective, but still, it is better to investigate the cross-transferability of SSRs in related genera to reduce the cost especially for underdeveloped labs (Kumar et al., 2014; Bharti et al., 2018). Evolutionary relatedness of the species affects the success of marker transferability. During cross-genera amplification of markers, 9 out of 10 primers could successfully be amplified, and merely one primer could not amplify in any of the related genera. To mitigate the amplification problem caused by genome complexity, PCR was carried out by changing the annealing temperature as suggested by Bharti et al. (2018).

The SSR transferability rate varied between 50% (Parsley) to 90% (celery). The frequency of clear and strong (scored as ++) bands varied from 20% (parsley) to 83.3% (fennel). A varied level of transferability of SSRs has been earlier reported in family Apiaceae (Cavagnaro et al., 2011; Kumar et al., 2014; Bharti et al., 2018). The reason for low to high transferability and amplicon quality (+or ++) may be due to (dis)similarity of dill seed genome with other studied crops in the current study.