Genome-wide characterization of two Aubrieta taxa: Aubrieta canescens subsp. canescens and Au. macrostyla (Brassicaceae).

Aubrieta canescens complex is divided into two subspecies, Au. canescens subsp. canescens, Au. canescens subsp. cilicica and a distinct species, Au. macrostyla, based on molecular phylogeny. We generated a draft assembly of Au. canescens subsp. canescens and Au. macrostyla using paired-end shotgun sequencing. This is the first attempt at genome characterization for the genus. In the presented study, ~165 and ~157 Mbp of the genomes of Au. canescens subsp. canescens and Au. macrostyla were assembled, respectively, and a total of 32 425 and 31 372 gene models were predicted in the genomes of the target taxa, respectively. We corroborated the phylogenomic affinity of taxa with some core Brassicaceae species (Clades A and B) including Arabis alpina. The orthology-based tree suggested that Aubrieta species differentiated from A. alpina 1.3-2.0 mya (million years ago). The genome-wide syntenic comparison of two Aubrieta taxa revealed that Au. canescens subsp. canescens (46 %) and Au. macrostyla (45 %) have an almost identical syntenic gene pair ratio. These novel genome assemblies are the first steps towards the chromosome-level assembly of Au. canescens and understanding the genome diversity within the genus.

Introduction

Brassicaceae is a moderately large, economically and scientifically important family comprising ca. 4000 species (Kiefer ; Koch ), including the model plant Arabidopsis thaliana, cabbage, turnip, radish, oil crops (e.g. Brassica, Raphanus and Camelina) and ornamentals (e.g. Aubrieta, Arabis, Hesperis, Lobularia and Matthiola) (Karl and Koch 2013; Nikolov ). This family has one of the highest speciation rates among terrestrial plant groups. Because polyploidy has a major impact on plant evolution, some diversification in Brassicaceae can be linked to whole-genome duplication (Hohmann ). Notably, the genus Draba has nine different ploidy levels, and the ratio of the number of taxa exhibiting polyploidy to the total number of taxa in the genus is 66 % (Jordon-Thaden and Koch 2008). Although three different ploidy levels were observed in the genus Arabis from the same tribe, Arabideae, the ratio of the number of taxa with polypoidy to the total number of taxa in the genus was 79 % (Koch ).

Aubrieta belongs to the tribe Arabideae in the expanded lineage II, with a chromosome number of 2n = 16 (Al-Shehbaz ; Beilstein ; Franzke ). Recently six clades (A, B, C, D, E, F) were identified for the family and the Brassicaceae lineage II was divided into three Clades B, C and D using the complete plastome (Couvreur ; Huang ; Guo ); phylogenomic analyses produced a well-resolved phylogenetic tree including six major clades in Brassicaceae (Nikolov ) but none of the Aubrieta taxa were included in these phylogenetic analyses.

The genus Aubrieta is difficult to classify taxonomically and is represented by 21 species that are distributed across the eastern Mediterranean region (Yüzbaşıoğlu ; Dönmez ; Koch ). The endemic species, Aubrieta canescens, has traditionally been divided into three subspecies: Au. canescens subsp. canescens, Au. canescens subsp. cilicica and Au. canescens subsp. macrostyla. Of these, Au. canescens is a moderately diverse species of the genus Aubrieta and is distributed across three phytogeographical regions of Turkey. Aubrieta canescens subsp. canescens is distributed in Central Anatolia and its adjacent regions, Au. canescens subsp. cilicica is mainly distributed in the Taurus Mountains and Au. canescens subsp. macrostyla is distributed in eastern Taurus. Although the morphological characteristics show only low differentiation between taxa, the habitats of the species are quite different, which may indicate a physiological rather than a morphological level of differentiation. Recent molecular studies have indicated that Au. canescens subsp. macrostyla should be separated as a distinct species (Koch ), and this taxonomic status is followed here. However, the Au. canescens complex requires further investigation to explore its phylogenetic relationship with the other taxa of the genus.

Members of the Brassicaceae family generally have small genomes, which allowed the sequencing of the first plant genome (Kaul ) and the sequencing of the highest number of genomes to date (plant genomes database, http://www.plabipd.de/plant_genomes_pa.ep). In addition, the development of genomic techniques has greatly facilitated research into the largest crucifer tribe, Arabideae, and the genome evolution of the seven subclades of the tribe has been inferred from centromere repositioning (Mandáková ). However, attempts to elucidate the genomic characters of the genus are limited and a publicly available genome assembly of Aubrieta has not yet been published.

Turkey is known to be one of the most important hotspots for native plants (Médail and Diadema 2009; Dönmez and Yerli 2018) and of genetic diversity centres of several crop plants (Zhukovsky 1951). All the complex taxa of Au. canescens are potentially ornamental plants and have been grown in numerous botanical gardens as rock plants.

In this study, we performed the whole-genome sequencing of Au. canescens subsp. canescens and Au. macrostyla, and their genome-wide characterization, including gene predictions, transposable element (TE) composition, variant identification and evolutionary comparisons.

Materials and Methods

Plant material, isolation and sequencing

Aubrieta canescens subsp. canescens (three accessions), and Au. macrostyla (three accessions) samples were used in this study. Seeds of the studied taxa were germinated in 2019, and one mature leaf originating from a single seed was selected for DNA extraction. Genomic DNA was extracted using a DNeasy Plant Pro Kit (Qiagen, Germany) and quantified using a Qubit fluorometer (Life Technologies, Foster City, CA, USA). Quality control was performed by analysing an aliquot of genomic DNA on an 1 % agarose gel. A paired-end genomic DNA library was constructed using the BGI DNB-seq platform and Macrogen company with a TruSeq DNA PCR-Free (350) kit. The library was run on NovaSeq, using a standard Illumina sequencing workflow [see].

Genome assembly

Before assembly, adapters and low-quality reads were removed using SOAPnuke v2.1.6, (Chen ) and quality control was performed using FastQCv0.11.9 (Andrew 2010). The optimum k-mer sizes were selected using the Kmerginie v1.7 software to match the reads correctly and rapidly (Chikhi and Medvedev 2014). Error correction of the raw data was performed using Karect v1.0 (Allam ). Genome assemblies were constructed using de novo assembly and iterative mapping approaches. The genome assemblies were generated de novo using the de Bruijn graph approach on the SPAdes v3.15 (Bankevich ) software with --careful and --only-assembler options. Subsequently, unmapped reads were iteratively mapped into the genome to correctly insert contigs that were mismatched or missing from other accessions using the BWA-mem v0.7.17 (Li 2013) algorithm. To create a less fragmented genome and close the gaps between contigs, we performed a post-assembly process using the Redundans v0.11 pipeline (Pryszcz and Gabaldón 2016). The assembly with the longest contig and the fewest number of contigs was selected according to the GAGE (GAGE: https://gage.cbcb.umd.edu/) criteria, for the rest of the downstream analyses.

Reference-assisted chromosome scaffolding

Given the availability of a relatively good reference genome (Arabis alpina latest version retrieved from http://www.arabis-alpina.org) for Au. canescens (n=8) and Au. macrostyla (n=8), a reference-assisted scaffolding approach was used to optimize the genome. The scaffolds were aligned to the A. alpina genome using blastn74 in the blast+ toolkit 2.8.0-alpha (Madden 2013). These alignments were used by chromosomer v0.1.3 fragmentmap command to perform the chromosome scaffolding (https://github.com/gtamazian/Chromosomer) (Tamazian ).

Assessment of assemblies

An assessment of the draft genomes was performed using QUAST v5.0.2, with the default parameters (Gurevich ). In addition, we investigated the completeness of genomes using the Benchmarking Universal Single-Copy Orthologs (BUSCO) v5.2.2 viridiplantae odb10 library (Simão ). Additionally, Minimap2 v2.22 was used to map the raw reads onto the final assemblies to evaluate the accuracy of the assemblies (Li 2018).

TEs and gene model prediction

To discover the de novo repeats in our assemblies, RepeatMasker (https://github.com/rmhubley/RepeatMasker) Repbase library and RepeatScout v1.0.5 software were used to predict TEs (Price ). Gene prediction was performed on Augustus-ab-initio v3.3.0, using BLAST hints (70 % threshold) and protein sequences of A. thaliana and A. alpina. The coding DNA sequences were extracted using getAnnoFasta Perl script of Augustus (Stanke ).

Variant investigation

Polymorphism was determined by scanning all the loci across the genome of the Au. canescens complex. To accomplish this, the draft assemblies were indexed using the BWA-mem algorithm. The clean data of each genome were then mapped to assembled files. Reads were sorted by removing PCR duplications using the Samtools v1.12 markdup command (Li ). The BAM files were indexed and prepared using the BamTools v2.5.2 index command to call the variants (Barnett ). Haplotype-based variants were generated from BAM files using Freebayes v1.3.5 (Garrison and Marth 2012). The parameters used were as follows: minimum coverage value, 15; minor allele frequency, 0.05; and minimum base quality score, 20.

Orthologs and synteny

A phylogenetic tree was constructed using the single-copy genes for Au. canescens subsp. canescens and Au. macrostyla and several other species representative of Clades A and B [see]. Homologous genes between species were analysed using OrthoFinder v2.5.4 (Emms and Kelly 2019). A single-locus species tree obtained from orthologous sequences was revealed by a pairwise sequence similarity approach using STAG and STRIDE algorithms prepared by OrthoFinder (Emms and Kelly 2017, 2018). To reconstruct the phylogeny, multiple sequence alignment was first performed using the MAFFT algorithm, followed by the construction of species trees using FastME in the OrthoFinder software. Diversification times were based on gene similarity calculations derived from the ancestral genes. We used the Satsuma v2.0 software to compare the syntenic genes of genomes with its closest relative, A. alpina (Grabherr ). We set the gene identity ratio to >70 % and the minimum coverage as 15×. After identifying the syntenic regions, we used Blastp and Blastn (NCBI-BLAST package, ftp://ftp.ncbi.nih.gov/blast/) to estimate the syntenic gene pairs in each genome. We used D-GENIES to visualize the assembled synteny (Cabanettes and Klopp 2018).

Results and Discussion

We analysed the raw reads from three accessions of Au. canescens subsp. canescens and three accessions of Au. macrostyla originating from wild populations of the taxa (Fig. 1A). Each assembly was iteratively mapped and compared, and high-quality genome assembly was included in the post-assembly analysis. In total, Au. canescens subsp. canescens samples had 30 Gb (~40×), and Au. macrostyla samples had 29 Gb (~40×) of paired-end sequencing reads [see]. All paired-end reads were used for the de novo assembly. Using the SPAdes software, 165 and 157 Mbp, respectively, of the draft genome of Au. canescens subsp. canescens and Au. macrostyla were assembled. The scaffold N50 sizes were 19.70 and 18.90 kbp and the longest contigs were 29 kbp and 28 kbp in Au. canescens subsp. canescens and Au. Macrostyla, respectively. BUSCO was employed to check the completeness of the draft assemblies, which detected 90 % of Au. canescens subsp. canescens, and 88 % of Au. macrostyla to be embryophyte genes in the two assemblies. In addition, short reads were mapped onto each corresponding assembly to evaluate the accuracy of the assemblies, and the mapping success was 98.4 % for Au. canescens subsp. canescens and 97.7 % for Au. macrostyla (Table 1; see).

Figure 1.

General habit and comparative gene density, repeat composition and polymorphism diversity across the genome of Aubrieta canescens subsp. canescens, and Au. macrostyla, respectively. (A) Flower and fruit morphology of the Au. canescens complex. (B) Gene density across the chromosomes. (C) Repeat composition. (D) Single nucleotide polymorphisms, multiple nucleotide polymorphisms, insertion, deletion and indel mutations across the genome.

Table 1.

Assembly statistics of Aubrieta canescens subsp. canescens and Au. macrostyla.

Assembly statistics	Au. canescens subsp. canescens	Au. macrostyla
Assembly strategies	Iterative and de novo	Iterative and de novo
Number of scaffolds	98	105
Longest scaffold (kb)	29	28
Guanine-cytosine content (%)	34.63	34.85
Mapping accuracy (%)	98.4	97.7
Complete BUSCOs percentage (%)	90	88
N50 (kb)	19.7	18.9
N75 (kb)	17.8	16.6
Assembled genome size (Mb)	165.0	156.9
Assembled contig numbers	818.200	647.738
Number of nucleotides per 100 kb	1862.51	1821.26

Gene model prediction and gene density

Gene model prediction was performed using the assembled scaffolds (Phred quality score > Q20 and >15× coverage). Aubrieta canescens subsp. canescens had a total of 32 425 coding DNA sequences, whereas Au. macrostyla had a total of 31 372 genes. A total of 417 sequences in Au. canescens subsp. canescens were found in the unplaced scaffolds of the chromosomes. Similarly, 393 sequences were found for Au. macrostyla. The highest gene density (20 % more than others) was observed on chromosome 8, followed by chromosome 4 in the two taxa (Fig. 1B; Table 2). Genomic profiling of the raw data was compared to evaluate heterozygosity, and the results showed that Au. macrostyla has a much higher genome heterozygosity (~2.5×), which is considered to involve more tandem sequences and gene models, than that of Au. canescens subsp. canescens (~0.4×, see; Tables 2 and 3) (Vurture ).

Table 2.

Chromosomal organization of gene models in Aubrieta canescens subsp. canescens and Au. macrostyla.

Chromosome	Gene model numbers of Au. canescens subsp. canescens	Gene model numbers of Au. macrostyla
Chr 1	3.989	3.875
Chr 2	2.892	2.698
Chr 3	3.960	4.033
Chr 4	4.503	4.295
Chr 5	3.441	3.306
Chr 6	3.400	3.242
Chr 7	3.971	3.686
Chr 8	5.852	5.844

Table 3.

Repetitive sequences in Aubrieta canescens subsp. canescens, and Au. macrostyla genome assembly.

	Repeat class	Repeat subclass	Repeat size of Au. canescens subsp. canescens (bp)	Repeat size of Au. macrostyla (bp)
Retrotransposons			11 644 448	9 586 050
	SINEs		446 655	440 701
	LINEs		2 915 226	2 741 343
		L1/CIN4	2 910 180	2 735 455
	Long terminal repeat elements		8 282 567	6 404 006
		Ty1/Copia	3 096 604	2 691 881
		Gypsy/DIRS1	4 957 195	3 510 223
DNA transposons			5 110 621	4 343 654
		hobo-Activator	1 240 611	1 130 175
		Tc1-IS630-Pogo	1 229 977	964 432
		Tourist/Harbinger	526 370	469 081
Small RNA			485 668	481 167
Satellite DNA			10 493	10 493
Simple sequence repeat (microsatellite)			2 825 779	2 602 091
Low-complexity DNA			965 076	894 604
Unclassified			529 526	428 768
Interspersed repeats			17 284 595	14 358 472
Total masked TE			21 112 636 bp (12.79 %)	17 896 883 bp (11.41 %)

Repeat composition

We analysed the repeat composition of the target Aubrieta taxa to infer the number of repetitive sequences in their genomes. In our whole-genome assembly, TEs comprised approximately 12.79 % and 11.41 % of the assembled parts of the genomes of Au. canescens subsp. canescens and Au. macrostyla, respectively. These sequences were divided into five major repeat classes: retroelements (55 % in Au. canescens subsp. canescens and 53 % in Au. macrostyla), DNA transposons (24 % in Au. canescens subsp. canescens and 24 % in Au. macrostyla), microsatellites (simple sequence repeats) (13 % in Au. canescens subsp. canescens and 15 % in Au. macrostyla) and low-complexity DNAs (5 % in Au. canescens subsp. canescens and 5 % in Au. macrostyla), and small RNAs (2 % in Au. canescens subsp. canescens and 3 % in Au. macrostyla) (Fig. 1C; Table 3; see).

Among these two taxa, the repeat content of the Au. canescens subsp. canescens genome was higher than that of Au. macrostyla genome (Fig. 1C). Specifically, the number of long terminal repeat elements was much higher in Au. canescens subsp. canescens. Among the classified repeat elements in the target taxa, Tourist/Harbinger made the lowest contribution to DNA transposons, followed by short interspersed nuclear elements (SINEs) in retrotransposons.

In terms of TEs, Au. canescens subsp. canescens had a higher percentage of class 1 transposon elements (34 %) than Au. macrostyla (32 %). Although the activity of long terminal repeat retrotransposons elements in the organism varies according to their interaction with the host genome, it can be significantly effective in the diversification of plants. The two taxa had the same percentage of class 2 transposons, Au. canescens subsp. canescens (38 %) and Au. macrostyla (38 %), whereas their length was 2 % and 7 % higher throughout the genome compared to A. thaliana (Kaul ). In Brassica oleracea, which shares a common ancestor with A. thaliana that diverged around 15–20 million years ago (mya) (Yang ), the proportion of mobile elements represents approximately 40 % of the entire genome (Chiu ; Jiang and Ramachandran 2013). A comparison of the whole-genome sequence of the Au. canescens with the high-quality assembly of A. alpina revealed that the related species, A. alpina, contains a higher percentage (approximately 25 %) of TEs (Willing ). Mobile elements play an important role in genomic and chromosomal evolution (Zhao ; Cheng ; Zhang ) and more recently several attempts have also been made to construct phylogenies using TE abundance as an informative character (Dodsworth ; Vitales ; Beric ). In addition, the repeat content is known to interfere with gene function and may result in the formation of variants that are responsible for phenotypic changes (Liu ). A comprehensive analysis of repetitive elements in our novel genome would provide valuable knowledge on morphological variations and evolution of the Au. canescens complex and its phylogenetic relationships in Aubrieta and Arabideae.

Variant identification

In the genome-wide variant exploration of the species, the minimum quality score of reads was 30.0 [seeand]. A total of 1 108 387 single nucleotide polymorphisms (SNPs) were found in Au. canescens subsp. canescens, representing 72 % of the variants found in the entire genome, and mostly being found on chromosomes 1 and 7. A total of 434 714 SNPs were identified in Au. macrostyla corresponding to 76 % of the total variants. In chromosome 2, rare but long base-pair variants were observed, whereas in chromosome 4 and chromosome 8, variants with dense but longer base pairs than the average distribution were detected [seeand]. In Au. canescens subsp. canescens the insertion/deletion ratio was 1.10, and the SNP transitions/transversions ratio was 1.39. In contrast, Au. macrostyla had an insertion/deletion ratio of 0.93 and a SNP transitions/transversions ratio of 1.42. In addition, the highest ratio of SNPs, MNPs, insertion, deletion and indel mutation content (14.2 % of all variants) was determined in Au. canescens subsp. canescens, whereas Au. macrostyla genome had the lowest variant content (Fig. 1D).

Evolutionary aspects

We compared gene data originating from the Aubrieta taxa with the related genome data of core Brassicaceae species (Fig. 2). Based on the phylogenomic tree, two representatives of Clade A, Camelina sativa and Arabidopsis lyrata, and members of Clade B, Brassica rapa species differentiated from A. thaliana, are well supported (Fig. 2A). Notably, the ratio of syntenic gene pairs (77–80 %) of B. rapa to that of A. thaliana and A. lyrata support this topology in previous studies (Cheng ). Moreover, the Au. canescens complex showed close affinity with A. alpina, a genus in Clade B. These results provide evidence that the Aubrieta genus must be classified as a member of Clade B.

Figure 2.

Genome evolution and comparative genomic analyses. (A) Homology-based phylogenetic tree of the Au. canescens complex and other Brassicaceae taxa. Gene numbers in nodes represent gene duplication events, and decimals in blue indicate node age. (B) Venn diagram showing the number of orthologous genes in Au. canescens subsp. canescens along with Arabis alpina, Brassica rapa, Camelina sativa, Arabidopsis thaliana, Arabidopsis lyrata and Au. macrostyla. (C) Relationships of syntenic genes between A. alpina and Au. canescens subsp. canescens, and Au. macrostyla. According to the identity scale, highly matched DNA sequences are indicated by dark green to light green dots (50–100 %), moderately matched sequences by orange dots (25–50 %) and poorly matched sequences by yellow dots (0–25 %). Sequences that did not match are shown in white.

Previously, Karl reported that 15 species of the genus Aubrieta diverged from the Arabideae tribes approximately 2.7–5.2 mya in the Eastern Mediterranean region. Likewise, orthologous genes indicate that Aubrieta species differentiated from A. alpina later, at 1.3–2.0 mya. Based on the homologous gene clusters obtained from OrthoFinder, we asserted that 2803 paralog genes distinguish A. alpina from the Aubrieta species. We consider that paralogs with a similar number of genes distinguish Au. macrostyla well from Au. canescens.

The percentage of assigned gene clusters in Au. canescens subsp. canescens and Au. macrostyla was 96.5 % and 95.4 %, respectively, within the 365 251 genes obtained from OrthoFinder. The number of gene clusters common to all species was 7078, whereas the number of single-copy gene clusters was 295. We consider that Au. macrostyla evolved before Au. canescens subsp. canescens and belongs to a different branch in terms of orthologs. The highest number of genes in a species-specific gene cluster was detected for Au. canescens subsp. canescens (244), followed by Au. macrostyla (232). Two taxa shared the most orthologous genes with perennial A. alpina (14 356) and A. lyrata (14 336) species (Fig. 2B).

Syntenic gene identification was performed between A. alpina (39 815 annotated proteins) and Au. canescens subsp. canescens (32 425) and Au. macrostyla (31 372). Prior to syntenic gene determination, duplicated tandem genes were removed from the genome. Arabis alpina returned 28 400 genes that showed synteny with 32 209 Au. canescens subsp. canescens and 28 982 Au. macrostyla genes. After eliminating redundant sequences, we identified syntenic gene pairs for each taxon. Although there were no remarkable differences between the species, the synteny ratio of A. alpina to Au. canescens subsp. canescens and Au. macrostyla was 46 % and 45 %, respectively. Most of the tandem arrays on the chromosomes 1, 3 and 8 of A. alpina showed strict synteny with the Au. canescens complex (Fig. 2C; Tables 4 and 5). These results show that chromosomes 4 and 7 are less informative in the syntenic gene arrangement, regardless of assembly size.

Table 4.

Syntenic orthologous genes between Arabis alpina to Aubrieta canescens and Au. macrostyla. *Number of overlapped sequences.

Syntenic orthologs to A. alpina	A. alpina vs. Au. canescens subsp. canescens	A. alpina vs. Au. macrostyla
Chr1 (arrays* | bp | genes)	377.339 | 7.194.247 | 1.926	424.768 | 7.540.313 | 1.788
Chr2 (arrays* | bp | genes)	485.449 | 5.466.368 | 1.414	423.693 | 4.799.368 | 1.230
Chr3 (arrays* | bp | genes)	296.393 | 7.323.310 | 1.853	274.384 | 6.922.344 | 1.797
Chr4 (arrays* | bp | genes)	628.093 | 7.719.562 | 1.544	610.820 | 7.587.290 | 1.563
Chr5 (arrays* | bp | genes)	416.086 | 5.828.293 | 1.470	390.924 | 5.661.399 | 1.599
Chr6 (arrays* | bp | genes)	303.688 | 5.863.068 | 1.707	275.829 | 5.454.359 | 1.597
Chr7 (arrays* | bp | genes)	459.382 | 6.696.666 | 1.524	443.276 | 6.433.374 | 1.436
Chr8 (arrays* | bp | genes)	676.307 | 10.542.223 | 1.717	647.019 | 10.509.024 | 1.815

Table 5.

Comparative homology of Arabis alpina and studied taxa Aubrieta canescens subsp. canescens, and Au. macrostyla.

Taxa	Number of genes	Number of removed tandem genes	Orthologous genes to A. alpina	Non-orthologs to A. alpina	Orthologous-based syntenic genes to A. alpina (identity >70 %)
Au. canescens subsp. canescens	32.425	216	19.373	13.052	13.155
Au. macrostyla	31.372	2.390	20.177	11.195	12.825

The ratio of orthologous-based syntenic genes of A. alpina to Au. canescens subsp. canescens (13 155/19 373) was 68 %, whereas that of A. alpina to Au. macrostyla (12 825/20 177) was 63.5 % (Table 5). We can consider the existence of these genes in additional arrays on chromosomes 1, 3, 6 and 8 in Au. canescens, whereas these genes were mostly ordered on chromosomes 1, 3 and 8 in Au. macrostyla (Table 4). Therefore, we can assume that Au. macrostyla has a clear genome-wide distinction, with orthologous domains that differ from those of A. alpina and Au. canescens.

The availability of genome data for Aubrieta taxa will enable us to analyse the population genomics within the genus Aubrieta in further studies. Additionally, this study will facilitate the elucidation of the Au. canescens complex, including Au. canescens subsp. cilicica. As a result, our data provide a significant contribution to the Aubrieta genome resource and novel insights into the evolution of Arabideae plants. The genome sequence produced here may help improve the potential horticultural value of Aubrieta taxa; however, a high-quality genome is needed for use in plant breeding programs and horticulture.

Supporting Information

The following additional information is available in the online version of this article—

Table S1. Sequencing reports of total six populations of Aubrieta canescens complex.

Table S2. Genomes used for comparative analysis in Brassicaceae family.

Table S3. Summary of repetitive elements in Aubrieta canescens subsp. canescens and Au. macrostyla genome.

Figure S1. BUSCO assessment for the taxa Aubrieta canescens subsp. canescens and Au. macrostyla.

Figure S2. Genomic profile plots of (A) Aubrieta canescens complex raw reads and (B) Au. macrostyla raw reads.

Figure S3. Quality score of bases obtained in variant identification. Selected bases are greater than 30 Phred score on all accessions.

Figure S4. Estimated coverage of bases used in variant identification. Minimum 15× base coverage was selected for determining polymorphisms on all accessions.

Figure S5. SNP density of Aubrieta canescens subsp. canescens across the chromosomes. The contig size of variants is kilobase pairs. Although the longest contigs were observed on chromosome 7, SNPs were observed in long contigs in chromosomes 3, 4 and 8 as well.

Figure S6. SNP density of Aubrieta macrostyla across the chromosomes. The contig size of variants is kilobase pairs. The longest contigs (30 kb) containing SNPs were observed on chromosome 2. Also, longer-than-average contigs (18 kb) were detected on chromosomes 4 and 7.

Click here for additional data file.