SVIM 可基于long reads(pacbio, ONT,HIFI)进行call SV,deletion, insertion, inversion, tandem duplications, interspersed duplication and translocations.
githup:https://github.com/eldariont/svim
文章:SVIM: structural variant identification using mapped long reads
1、安装
conda create -n svim_env --channel bioconda svim
2、简单操练
所需数据:
- long reads FASTA/FASTQ (压缩或uncompressed均可)
- or long reads比对得到的bam文件, 需要
在进行long reads比对的时候,可以选择 NGMLR
或者minimap2均可以。
SVIM主要包括4个步骤:
- collect: 基于Long reads 检测SV
- cluster:相同的SV进行合并
- combine:来自不同基因组区域的cluster进行合并
- genotype:确定基因型
svim alignment my_sample my_alignments.bam my_genome.fa
输出文件:
- The log file: SVIM_{date}_{time}.log
- The SV calls in VCF format: variants.vcf
- The SV calls in BED format: candidates/candidates_*.bed
- Intermediate signature clusters in BED format: signatures/*.bed
当然也可以进行适当的过滤,比如
## score > 10
bcftools view -i 'QUAL >= 10' variants.vcf'.