NAME

Bio-ToolBox-Nucleosome - Scripts for working with nucleosome sequences

DESCRIPTION

This is a collections of specialized scripts written for working with next-generation sequencing of MNase-digested mononucleosome DNA (MNase-Seq). At the time these were written many years ago (late 00s), there was little else written for identifying, mapping, and working with nucleosomal reads. There are now a few other projects that do so. This package may not be the most accurate, but it works reasonably well, at least for S. cerevisiae.

These were initially part of the Bio::ToolBox package, at least in early versions (earlier than 1.30). For a while, they were part of the legacy Bio::ToolBox::Legacy package, but have now moved out on its own.

USAGE

These scripts were written primarily for S. cerevisiae nucleosomes. They may or may not work with other model system. Since yeast have very well-ordered nucleosomes, it's relatively easy to map them. This package has primarily four scripts.

• map_nucleosomes.pl

  This is the main script to map nucleosome positions. It maps based on the position of the peak of nucleosome fragment midpoints within an incrementally sliding window. Search windows are based relavent to the previously identified nucleosome, in accordance to nucleosome stacking laws. If you can pick out nucleosome positions just by eye from nucleosomal fragment midpoint data in a genome browser without fancy smoothing or hidden Markov model states, then this program will work. This, of course, presumes a good degree of duplicate fragment positions (stack of identical nucleosomes) which is common with MNaseSeq data, but PCR-duplication is a concern. You may wish to threshold duplicate percentages between replicates and/or conditions using bam_partial_dedup from the MultiRepMacsChIPSeq package.

  Generate bigWig files of nucleosome midpoint data using, for example, bam2wig.pl. Either single- or paired-end alignments could and have been used successfully.

  UPDATE: Actually, working with "skinny" nucleosome coverage appears to give the best results. Generate nucleosome coverage over the central 50% (or less) portion of the nucleosome. This "accentuates" the nucleosome occupancy curves while averaging out the density of nucleosome midpoints. There are much fewer "off-center" nucleosomes, fewer overlaps, and fewer "weak" nucleosomes called (because you can increase stringency) with this technique. Generate skinny nucleosome coverage from single-end alignments with bam2wig.pl:

    bam2wig.pl --extend --extval 75 --shift --shiftval 37 --rpm --bin 1 --bw --in file.bam
  

  With paired-end data, use center-span:

    bam2wig.pl --pe --cspan --extval 75 -rpm --bin 1 --bw --in file.bam
  

• verify_nucleosome_mapping.pl

  This is an optional script, but will verify how well the nucleosomes were mapped with the script above, and optionally re-adjust and/or discard nucleosomes. Probably a good idea to do so.

• get_actual_nuc_sizes.pl

  A script to take the mapped nucleosomes, grab the actual paired-end alignments from the original Bam file, and try to calculate the actual nucleosome size. YMMV

• intersect_nucs.pl

  A script to take two list files of mapped nucleosomes, and intersect them, identifying which nucleosomes overlap and calculating some rudimentary statistics. May be marginally useful....

• correlate_position_data.pl

  This script is not part of this package but rather part of the Bio::ToolBox package and is particularly useful here when working with nucleosomal data between different conditions. It will identify differences in the spatial position of signal over defined intervals, such as mapped nucleosomes. Differences between two datasets can be reported with a p-value, and importantly, an optimal shift in the signal. In other words, if the nucleosome signal in a mutant condition has shifted relavent to wildtype, this program will identify it, with caveats to the original mapping, nucleosome fuzziness, etc.

Identifying positioned nucleosomes in yeast

The subfolder yeast_positioned_nucleosomes contains a bash script for using a number of BioToolBox programs for identifying positioned nucleosomes, i.e. the plus1, plus2, plus3, etc nucleosomes over each transcript. It may be useful to you, either as is or with modifications.

REQUIREMENTS

These are Perl modules and scripts. They require Perl and a unix-like command-line environment. They have been developed and tested on Mac OS X and linux; Microsoft Windows compatability is not tested nor guaranteed.

These scripts require the installation of Bio::ToolBox, specifically version 1.69, and all the requirements therein, particularly Bam and bigWig file adapters. There are installation instructions on the BioToolBox page.

This will also require the Bio::ToolBox::Legacy module as well. NOTE: This requires Bio::ToolBox version 1.69; it is incompatible with the latest versions.

INSTALLATION

Installation is simple with the standard Perl incantation. However, it's highly recommended to install prerequisites and dependencies first. See the Bio::ToolBox installation page.

perl ./Build.PL
./Build
./Build install

AUTHOR

Timothy J. Parnell, PhD
Huntsman Cancer Institute
University of Utah
Salt Lake City, UT, 84112

LICENSE

This package is free software; you can redistribute it and/or modify it under the terms of the Artistic License 2.0. For details, see the full text of the license in the file LICENSE.

This package is distributed in the hope that it will be useful, but it is provided "as is" and without any express or implied warranties. For details, see the full text of the license in the file LICENSE.