【Juicer loop】HiCCUPS
企業(yè)微信截圖_c04d145e-7d8f-4626-ae7c-109264a23a2d.png
HiCCUPS call loop的算法如上。
用法如下：
hiccups [-m matrixSize] [-c chromosome(s)] [-r resolution(s)]
[-k normalization (NONE/VC/VC_SQRT/KR)] [-f fdr]
[-p peak width] [-i window] [-t thresholds]
[-d centroid distances] <HiC file> <outputDirectory> [specified_loop_list]
重要參數(shù)說(shuō)明：
-r ：指定call loop 的resolution，多個(gè)分辨率以 , 分割 如5000,10000 注意最高分辨率為5000,默認(rèn)5000,10000, 25000
HiC file ：為輸入文件即.hic files
outputDirectory：輸出結(jié)果目錄
-f ：fdr 閾值 多個(gè)閾值要用逗號(hào)分割如：“- f 0.1,0.15" fdr默認(rèn)為 10%
簡(jiǎn)單示例:
java -jar juicer_tools.1.7.5_linux_x64_jcuda.0.8.jar hiccups GM.chr10.hic GM/chr10/
集群實(shí)例:
java -Xmx25000m -Xms8000m -jar juicer_tools.1.7.5_linux_x64_jcuda.0.8.jar hiccups -r 5000 sam1.chr10.hic /sam1/chr10
輸出結(jié)果目錄：
- enriched_pixels_5000
- fdr_thresholds_5000
- merged_loops
- postprocessed_pixels_5000
重要結(jié)果文件說(shuō)明:
merged loop
merged loop文件共有20列，如下所示
chr1 x1 x2 chr2 y1 y2 color observed expectedBL expectedDonut expectedH expectedV fdrBL fdr Donut fdrH fdrV numCollapsed centroid1 centroid2 radius
詳細(xì)說(shuō)明如下：
- chromosome = the chromosome that the loop is located on
- x1,x2 = the coordinates of the upstream locus corresponding to the peak pixel
- y1,y2 = the coordinates of the downstream locus corresponding to the peak pixel
- color = the color that the feature will be rendered as if loaded in Juicebox
- observed = the raw observed counts at the peak pixel
- expected_[bottom_left, donut, horizontal, vertical] = the expected counts calculated using the [bottom_left, donut, horizontal, vertical] filter
- fdr_[bottom_left, donut, horizontal, vertical] = the q-value of the loop calculated using the [bottom_left, donut, horizontal, vertical] filter
- number_collapsed = the number of pixels that were clustered together as part of the loop call
- centroid1 = the upstream coordinate of the centroid of the cluster of pixels corresponding to the loop
- centroid2 = the downstream coordinate of the centroid of the cluster of pixels corresponding to the loop
- radius = the Euclidean distance from the centroid of the cluster of pixels to the farthest pixel in the cluster of pixels
