juice tools_LSF - BiocottonHub/BioSoftware GitHub Wiki
使用软件中提供的LSF集群脚本 cd 仓库/LSF
tree -L 1
.
├── hg19_chromsome.bed #染色体长度信息
├── README.md
├── references #参考基因文件夹
├── restriction_sites #酶切文件夹
├── scripts #LSF脚本文件夹
└── test ##Hi-C 原始数据文件夹
在test/fastq
文件夹中存放原始的测序数据
在scripts
文件夹中存放LSF脚本
由于软件需要在GPU节点进行计算,需要使用到cuda软件;首先配置自己的环境变量
cuda='/usr/local/cuda/'
export PATH=${cuda}"bin:"$PATH
export LD_LIBRARY_PATH=${cuda}"lib:"${LD_LIBRARY_PATH}
-
test/aligned/inter_30.hic
文件是使用juicer计算得到的 - scripts目录是对应
LSF
版本的软连接 - 指定计算节点为GPU节点
gpu
- 单分辨率运行
bsub -q gpu -J juicer -gpu num=1 -n 10 -e juicer.err -o juicer.out -R span[hosts=1] "java -jar scripts/juicer_tools.jar hiccups -m 512 -r 10000 -f 0.1 -p 2 -i 5 -d 20000 -c 1 -k KR --threads 10 --ignore-sparsity test/aligned/inter_30.hic ./chrosomeLoop"
- 多分辨率运行
bsub -q gpu -J juicer -gpu num=1 -n 10 -e juicer.err -o juicer.out -R span[hosts=1] "java -jar scripts/juicer_tools.jar hiccups -m 512 -r 5000,10000 -f 0.1,0.1 -p 4,2 -i 7,5 -d 20000,20000 -c 1 -k KR --threads 10 --ignore-sparsity test/aligned/inter_30.hic ./chrosomeLoop"
在单个分辨率下的计算时,在输出目录中仅仅只包含单个文件
fdr_thresholds_10000
多分辨率情况下输出结果
.
├── fdr_thresholds_10000
└── fdr_thresholds_5000
划分compartments
- 划分chr1号染色体在1M的分辨率下,得分
java -jar scripts/juicer_tools.jar eigenvector KR test/aligned/inter_30.hic chr1 BP 1000000 >111
结果
- 每一行代表1M的区间的得分
0.010239605229520426
-0.0726968084857949
-0.04291854477085067
-0.029418479599141876
-0.004632664168007863
0.015414336275401206
-0.08943360513489482
-0.06504310611117355