GenomeDISCO

评估从染色体捕获实验获得的一对接触图的相似性。GenomeDISCO设计用于评估从染色体构象捕获实验获得的染色质接触图的一致性和可重复性。它对测序深度，结点和边缘缺失噪声，结构域边界的变化以及距离依赖性的细微差别的识别灵敏度高，将生物重复与不同细胞类型区分开来。

GenomeDISCO (DIfferences between Smoothed COntact maps 平滑接触图谱间的差异) is a package for comparing contact maps of 3D genome structures, obtained from experiments such as Hi-C, Capture-C, ChIA-PET, HiChip, etc. 在比较接触图之前，它在接触图图谱上随机行走进行平滑，从而产生一致性分数，可用于生物重复的质量控制。

比较两个接触图

1. 得到配置文件

/genomedisco/examples/configure_example.sh

2. 进行一致性分析

cd genomedisco
genomedisco run_all --metadata_samples examples/metadata.samples --metadata_pairs examples/metadata.pairs --bins examples/Bins.w50000.bed.gz --outdir examples/output 

Inputs

运行  GenomeDISCO 前，

要有以下文件

• contact map

•  For each of your samples, you need a file containing the counts assigned to each pair of bins in your contact map,and should have the format chr1 bin1 chr2 bin2 value.

• 对你的每个样本，你需要一个 包含分配到你接触图谱每对bins的数量的文件，也得有有染色体和其bin值的这样的格式的内容

• Note: GenomeDISCO assumes that this file contains the contacts for all chromosomes, and will split it into individual files for each chromosome

• 注意DISCO假设认为这个文件包含了所有染色体的接触，并将把这个文件拆成对每个染色体的单个文件。

• bins 

• This file contains the full set of genomic regions associated with your contact maps, in the format chr start end name where name is the name of the bin as used in the contact map files above. GenomeDISCO supports both fixed-size bins and variable-sized bins (e.g. obtained by partitioning the genome into restriction fragments).

• 该文件包括了和你的接触图谱相关的一整套基因组区域，格式是染色体 开始 结束 名字
• 名字是上面说的用于接触图谱的bin的的名字，DISCO同时支持固定大小和可变大小的bins值
• 例如其中一些bins是通过把基因组切分成限制性片段获得的

GenomeDISCO 需要下列输入文件:

• --metadata_samples 要比较的样本信息.                   Tab-delimited file, with columns "samplename", "samplefile". Note: each samplename should be unique. Each samplefile listed here should follow the format "chr1 bin1 chr2 bin2 value

• --metadata_pairs    每一行是一对要比较的样本名。   in the format "samplename1 samplename2". Important: sample names used here need to correspond to the first column of the --metadata_samples file. 

• --bins A (gzipped)   分析中用到的所有bins的bed file。 It should have 4 columns: "chr start end name", where the name of the bin corresponds to the bins used in the contact maps.

• --re_fragments  如果基因组中的bins不均一 就加上flag  (例如，如果它们是基于限制性内切酶的).默认情况下这些代码认为bins是均一的长度。

• --parameters_file  生物重复和QC分析的参数   For details see "Parameters file"

• --outdir       Name of output directory. DEFAULT: replicateQC

• --running_mode        The mode in which to run the analysis. This allows you to choose whether the analysis will be run as is, or submitted as a job through sge or slurm. Available options are: "NA" (default, no jobs are submitted). Coming soon: "sge", "slurm"

• --concise_analysis Set this flag to obtain a concise analysis, which means replicateQC is measured but plots that might be more time/memory consuming are not created. This is useful for quick testing or running large-scale analyses on hundreds of comparisons.

• --subset_chromosomes Comma-delimited list of chromosomes for which you want to run the analysis.该参数选择部分染色体进行分析。 默认在所有染色体上进行分析。This is useful for quick testing

　

　

来源：

https://omictools.com/genomedisco-tool

https://github.com/kundajelab/genomedisco