NucleoATAC

Update 2024/08/26

This is a fork of the NucleoATAC package authored by Alicia Schep in the Greenleaf Lab, a package for calling nucleosome occupancy from ATAC-seq data. We modify the package to take fragments files as inputs instead of BAM files. This fork is not being officially maintained. Please refer to the original package for the last stable version.

Installation

Create a conda environment

# loading the module first seems to be necessary for properly loading cython
module load python/2.7.13
conda create -n "nucleoatac" python=2.7.13
conda activate nucleoatac

Install the package

cd NucleoATAC
pip install --editable .

NOTE: to edit the package and install the new version, update the version number in setup.py for documentation, then re-run the above pip command. Test the install with pyatac --version.

Updates to functionality

NOTE: nucleoatac run doesn't work with the modifications for fragment sizes yet. Run commands individually as per the docs.

Allowing fragments files as input

Goal: allow nucleosome calling to be performed with tabix-indexed compressed fragments files in BED format
The pyatac.fragmentsizes module has been updated so that the FragmentSizes class can now accept a fragments file instead of a BAM file for calculation of the fragment size distribution. This makes use of new functions in the nucleoatac.fragments_handling module.
The following commands now can now accept a fragments file instead of BAM file:
- pyatac sizes
- nucleoatac occ
- nucleoatac nuc
The pyatac.chunkmat2d.FragmentMat2D module has been updated so that the FragmentMat2D class can now accept a fragments file instead of a BAM file for calculation of the fragment size distribution, which will fetch the chunk region from the fragments file using pysam's tabix interface, and then populate the 2D matrix
Add chunk handling for the fragemnts file as well
When commands required BAM files for calculation of fragment sizes, the fragment sizes output from previous commands is now required instead

Restrict analysis to chromosomes of interest

Goal: calculate the fragment size distribution on the whole dataset, but restrict the occupancy calculation to a subset of chromosomes to allow users to debug or further parallelize the analysis
The nucleotac occ command now accepts a --chroms_keep argument that allows users to specify a list of chromosomes to restrict the occupancy calculation to, and a chunk list will only be generated for these chromosomes.
This can be used to manually parallelize the runs, or do a quick run for debugging

Misc

nucleoatac occ and pyatac sizes changed to produce .pdf instead of .eps files
More verbose messaging added in nucleoatac occ and nucleoatac nuc

Timing

For reference, running only on chr1 (with 7,267,633 fragments) on 4 cores and 64G of memory took about 14 minutes.

Command run:  /path/to/.local/bin/nucleoatac occ --fragments ../data/Eye_c11__sorted.tsv.gz --bed ../data/Eye_c11__peaks_overlap_filtered.narrowPeak --fasta
 /path/to/GRCh38_no_alt_analysis_set_GCA_000001405.15.fasta --out ../out/04-Eye_c11__chr1 --cores 4 --chroms_keep chr
1
nucleoatac version 0.3.4
start run at: 2024-08-27 14:48
---------Computing Occupancy and Nucleosomal Insert Distribution----------------
@ NOTE: restricting analysis to chromosomes: chr1
@ calculating fragment sizes...
@ fitting fragment size distribution...
@ calculating occupancy...
@ compressing and indexing output files...
@ making figure
@ done.
end run at: 2024-08-27 15:02

Previously

This package is no longer being actively maintained; feel free to post issues that others in the community may respond to, but this package will likely not be updated further. Additionally, if anyone wants to maintain a fork of the package or has developed an alternative package for similar purposes, would be happy to link to that repo here.

Python package for calling nucleosomes using ATAC-seq data. Also includes general scripts for working with paired-end ATAC-seq data (or potentially other paired-end data).

Please cite our paper at Genome Research if you use this tool in your research.

Please use GitHub Issues to bring up any errors that occur with software rather than emailing authors.

Note on Versions:

version 0 represents code used for biorxiv manuscript
version 0.2.1 was used for Genome Research manuscript (See Supplemental Information as well)

Documentation can be found at http://nucleoatac.readthedocs.org/en/latest/

If you want to easily read in NucleoATAC outputs into R for further processing or exploration, check out NucleoATACR

Currently NucleoATAC only supports Python 2.7 (No Python 3).

sjessa/NucleoATAC