【3】SALSA-基于HIC数据辅助组装长读长组装结果

1. 安装编译

git clone https://github.com/marbl/SALSA.git
cd SALSA
make -j8

##SALSA2依赖环境: Python 2.7，Networkx version < 1.2，BOOST libraries

2. 文件输入：

三代组装结果: contigs.fasta

Hi-C比对后BAM: alignment.bam

bwa mem -SP5M -t 20 contigs.fasta HICLib_R1.fastq.gz HICLib_R2.fastq.gz 
     | samtools view -hF 256 - 
     | samtools sort -@ 20 -o alignment.bam -T tmp.ali

samtools index alignment.bam

(可选) GFA文件: contigs_graph.gfa

NOTE: SALSA2从PacBio/ONT的组装结果开始，建议提供记录模棱两可重建位置的GFA文件。Hi-C数据回贴到contig上，contig根据Hi-C的覆盖度分析潜在的组装错误，然后被打断。之后根据GFA和Hi-C信息构建hybrid scaffold graph，最后scaffold根据该graph进行重建。

3. 运行

#1.convert bam to bed file
bamToBed -i alignment.bam > bwa.bam.bed
#2.sort bed file
sort -k 4 bwa.bam.bed > tmp && mv tmp bwa.bam.bed
#3.build length file of contig genome
samtools faidx contigs.fasta
#激活python 2.7环境
conda activate py2.7
/home/user/plants/02.hic_assembly/004.asm_by_SALSA/SALSA/run_pipeline.py -a contigs.fasta -l contigs.fasta.fai -b bwa.bam.bed -e AAGCTT -o scaffolds -m yes

NOTE: 

-m yes #使用Hi-C对输入assembly进行纠错；

-e {Your Enzyme} #根据实际情况修改，如果是MboI就写-e GATC，HindIII就写-e AAGCTT；

最终的输出结果在-o指定的目录下，其中scaffolds_FINAL.fasta和scaffolds_FINAL.agp是两个最主要的结果，其中scaffolds_iteration_x.p里的x取决于你的迭代次数。

亲测：使用SALSA基于HIC数据组装后，scaffold N50提高2～3倍。（仅供参考）​​​​​​​

SALSA参数：（https://github.com/marbl/SALSA）

usage: run_pipeline.py [-h] -a ASSEMBLY -l LENGTH -b BED [-o OUTPUT]
                       [-c CUTOFF] [-g GFA] [-e ENZYME] [-i ITER] [-x DUP]
                       [-s EXP] [-m CLEAN] [-f FILTER] [-p PRNT]

SALSA Iterative Pipeline

optional arguments:
  -h, --help            show this help message and exit
  -a ASSEMBLY, --assembly ASSEMBLY
                        Path to initial assembly
  -l LENGTH, --length LENGTH
                        Length of contigs at start
  -b BED, --bed BED     Bed file of alignments sorted by read names
  -o OUTPUT, --output OUTPUT
                        Output directory to put results
  -c CUTOFF, --cutoff CUTOFF
                        Minimum contig length to scaffold, default=1000
  -g GFA, --gfa GFA     GFA file for assembly
  -e ENZYME, --enzyme ENZYME
                        Restriction Enzyme used for experiment
  -i ITER, --iter ITER  Number of iterations to run, default = 3
  -x DUP, --dup DUP     File containing duplicated contig information
  -s EXP, --exp EXP     Expected Genome size of the assembled genome
  -m CLEAN, --clean CLEAN
                        Set this option to "yes" if you want to find
                        misassemblies in input assembly
  -f FILTER, --filter FILTER
                        Filter bed file for contigs present in the assembly
  -p PRNT, --prnt PRNT  Set this option to "yes" if you want to output the
                        scaffolds sequence and agp file for each iteration

其他参考：SALSA:基于Hi-C辅助组装长读长组装结果_xuzhougeng blog-CSDN博客

敲黑板：SALSA无法得到染色体级别的输出结果，因此得到的结果需要再借助于遗传图谱或者近缘参考基因组或者HIC图谱挂载到染色体级别。

见下一推送的RagTag相关内容。

【4】RagTag-基于近缘/同物种的基因组同源组装_AlisLee的博客-CSDN博客