Planned intervention: On Wednesday June 26th 05:30 UTC Zenodo will be unavailable for 10-20 minutes to perform a storage cluster upgrade.
Published April 6, 2020 | Version v5
Dataset Open

Chromosight benchmarks and processed data

Description

Input data and scripts required to rerun chromosight's benchmarks and associated figures from the manuscript, as well as output results in text format.

The record contains 4 tarballs corresponding to two different benchmarks and various processed files. Each tarball contains the scripts, input and output data of its benchmark:

Performance benchmark:

This benchmark is contained in "20200406_benchmark_chromosight_performance.tar.gz" and compares chromosight running time and RAM use with 2 other softwares. This benchmark is run on a real high resolution human Hi-C matrix with different subsampling values. Benchmark scripts are expected to be run on a regular laptop or desktop.

Results benchmark:

This benchmark is contained in "20200406_benchmark_chromosight_results.tar.gz" and assess chromosight's ability to detect chromatin loop patterns on Hi-C contact maps. Chromosight is compared to 4 other softwares. For each software, precision, recall (=sensitivity) and F1 scores are measured using 2000 small synthetic Hi-C matrices with known loop coordinates. Each software is run with a range of 50-200 parameter combinations for all data. Scripts to run this benchmark are written to run as a job array on a SLURM computing cluster to reduce compute time.

Processed data files

Intermediate and output files used throughout the manuscript. This includes contact matrices in cool format, genomic intervals in BED or BED2D format and loop calls from public data for different softwares.

 

Simulation input:

The file "simulation_input.tar.gz" contains the inputs used to generate synthetic Hi-C contact maps used in the results benchmark. it consists of a contact map from the chromosome 5 Saccharomyces cerevisiae strain W303 from this project, and a corresponding set of domain border coordinates detected on this contact map as described in the methods of Chromosight's paper. Border coordinates are counted in bins from the start of chromosome 5. Both files can be fed to the "chromo_simul.py" script available on github as follows to generate synthetic contact maps:
`python chromo_simul.py matrix.cool chr5 borders.txt out_dir`

Files

Files (2.6 GB)

Name Size Download all
md5:7daf25490681ce9e32f55f343c263d00
317.9 MB Download
md5:3d00efc58a14bff3d3a92355a5a0de79
2.0 GB Download
md5:4c16e8b599d82f4eb362e262caa5f9c7
243.7 MB Download
md5:621ea03205a3c1e59aa8060899e3f17a
37.8 kB Download

Additional details

Related works

Is supplement to
Preprint: 10.1101/2020.03.08.981910 (DOI)