docker-4dn-insulation-scores-and-boundaries-caller
This repo contains the source files for a docker image stored in docker-4dn-insulation-scores-and-boundaries-caller.
Table of contents
- Cloning the repo
- Tool specifications
- Building docker image
- Benchmarking tools
- Sample data
- Tool wrappers
Cloning the repo
git clone https://github.com/4dn-dcic/docker-4dn-insulation-scores-and-boundaries-caller
cd docker-4dn-insulation-scores-and-boundaries-caller
Tool specifications
Major software tools used inside the docker container are downloaded by the script downloads.sh. This script also creates a symlink to a version-independent folder for each software tool. In order to build an updated docker image with a new version of the tools, ideally only downloads.sh should be modified, but not Dockerfile, unless the new tool requires a specific APT tool that need to be downloaded.
The downloads.sh file also contains comment lines that specifies the name and version of individual software tools.
Building docker image
4dn-insulation-scores-and-boundaries-caller You need docker daemon to rebuild the docker image. If you want to push it to a different docker repo, replace 4dndcic/4dn-insulation-scores-and-boundaries-caller with your desired docker repo name. You need permission to push to 4dndcic/4dn-insulation-scores-and-boundaries-caller.
docker build -t 4dndcic/4dn-insulation-scores-and-boundaries-caller .
docker push 4dndcic/4dn-insulation-scores-and-boundaries-caller
You can skip this if you want to use an already built image on docker hub (image name 4dndcic/4dn-insulation-scores-and-boundaries-caller). The command 'docker run' (below) automatically pulls the image from docker hub.
Benchmarking tools
To obtain run time and max mem stats, use usr/bin/time that is installed in the docker container. For example, run the following to benchmark du.
docker run 4dndcic/4dn-insulator-score-caller /usr/bin/time du 2> log
cat log
The output looks as follows:
0.02user 0.82system 0:00.87elapsed 96%CPU (0avgtext+0avgdata 2024maxresident)k
0inputs+0outputs (0major+103minor)pagefaults 0swaps
The benchmarking result goes to STDERR, which can be collected by a file by redirecting with 2>.
Maxmem is 2024KB in this case ('maxresident'). Run time is 0.87 second. ('elapsed')
Sample data
Sample data files that can be used for testing the tools are included in the sample_data folder. These data are not included in the docker image.
Tool wrappers
Tool wrappers are under the scripts directory and follow naming conventions run-xx.sh. These wrappers are copied to the docker image at built time and may be used as a single step in a workflow.
# default
docker run 4dndcic/4dn-insulation-scores-and-boundaries-caller
# specific run command
docker run 4dndcic/4dn-insulation-scores-and-boundaries-caller <run-xx.sh> <arg1> <arg2> ...
# may need -v option to mount data file/folder if they are used as arguments.
docker run -v /data1/:/d1/:rw -v /data2/:/d2/:rw 4dndcic/4dn-insulation-scores-and-boundaries-caller <run-xx.sh> /d1/file1 /d2/file2 ...
run-insulation-scores-and-boundaries-caller.sh
This calls the 'get_insulation_scores_and_boundaries.py' file which gets the diamond insulation scores and boundaries from mcool files produced by Hi-C data and outputs them in bigwig and bed format respectively.
- Input: a mcool file
- Output: a bigwig file with insulation scores and a bed file with boundaries
Usage
Runs the following in the container
run-insulation-scores-and-boundaries-callerr.sh <mcoolfile> <outdir> <filename> --binsize --windowsize --bweak --bstrong --cutoff --pixels_frac
# mcool file: input mcool
# outdir: output directory
# binsize: the size of the bins
# windowsize: the size of the sliding diamond window
# bweak: the threshold for defining weak boundaries
# bstrong: the threshold for defining strong boundaries
# cutoff: minimal distance allowed to a bad bin. Do not calculate insulation scores for bad bins closer to that distance
# pixels_frac: The minimal fraction of valid pixels in a diamond to be used in boundary picking and prominence calculation