Genomes resequencing of CMS line ***** and fertility restorer line *****
Sample collection and SLAF library preparation
Fresh leaves were obtained from the parent lines and F2, frozen with liquid nitrogen, extracted by the CTAB method, and assessed for the quality of DNA by 1% agarose gel electrophoresis. The purity of DNA was examined using the NanoPhotometer® spectrophotometer (Implen, CA, USA). The DNA concentration was estimated using Qubit® DNA Assay Kit in Qubit® 2.0 Fluorometer (Life Technologies, CA, USA).
We used 1.5 μg DNA/sample as input material for the preparations of the sample. We have chosen to use RsaIas restriction enzyme in the electronic enzyme-digestion projections to the reference genome sequences of cotton. Sequencing libraries were generated using RsaIof restriction enzyme according to the manufacturer’s recommendations, and index codes were added to ascribe the sequences to each sample. Briefly, the DNA sample was fragmented by sonication to a size of 350 bp. Then, the DNA fragments were end-polished, A-tailed, and ligated with the full-length adapter for Illumina sequencing by PCR amplification. Consequently, the PCR products were purified (Agencount® AMPure® XP, USA), and libraries were analyzed for size distribution by Agilent2100 Bioanalyzer and quantified by real-time PCR.
Illumina sequencing
The libraries constructed above were sequenced by Illumina HiSeq ™2500 (Illumina, Inc., San Diego, USA) platform at Genome and Biomedical Sciences Facility of  University of California, Davis, CA (http://gbsf.ucdavis.edu/) and 150 bp paired-end reads were generated with an insert size approximately 350***** bp.
Data analysis, data filtering, and alignment
Brassica_napus_v4.1.chromosomes.fa
The recently released genome of Brassica napus was downloaded from Cotton Research Institute (CRI) of Nanjing Agricultural University in China. (http://mascotton.njau.edu.cn/Data.htm, v1.1) and used as a reference genome [19]. FastQC-toolkit (v 0.11.2) and Trimmomatic (v0.33) were used to filter out the low-quality reads in the following order: (i) reads with the adapter; (ii) low quality or N bases (below quality 3); (iii) reads with the average quality per base drops below 15 for a 4-base wide sliding window; (iv) reads which are less than 36 bases long after these steps.  The remaining clean reads were aligned to the reference Brassica napus genome using BWA-MEM (v0.7.16a) [20] and default parameters. Sequence Alignment/Map tools (SAMtools) (v1.4.1) [21] was applied to sort and index the resulting binary alignment map (BAM) format files. The duplicates were excluded using SAMtools (v1.102), and the final sorted bam files were utilized in the downstream analysis. Variant calling and filtering were performed in order to reduce the inaccuracy of the alignment. Variant calling was conducted using freebayes Tools (v1.0.2). [2223]. The variants that fulfilled the following criteria were filtereded (1) sites with indels; (2) min alleles < 2; (3) sites with Quality value below 70; (5) genotypes with a quality <30  (5) sites mean depth < 20 or >500. 
Moreover, the variants were filtered further if the coverage was <10, the cluster SNPs were >2 in a 5 bp window, if the SNP around the Indel was within 5 bp. SV detection and annotation BreakDancer was used to predict the five types of structural variants (SVs): insertions (INSs), deletions (DELs), inversions (INVs), intra-chromosomal translocations (ITXs), and inter-chromosomal translocations (CTXs) from next-generation paired-end sequencing reads utilizing the read pairs mapped with excessive separation distances or orientation. The SVs with read depth < 2 were filtered. Bedtools was employed to annotate the detected DELs, INSs, and INVs. The detection and annotation of CNVs (copy number variations) refers to a normal variation in the number of copies of ≥1 sections of some genomic fragments. We used CNVnator (parameters: -call 100) for the identification of CNVs and bedtools for annotations.