Python package that helps intersect BED files with sequence conservation tracks.

This package was inspired by the need to extract conservation scores for a given set of BED intervals. It is based on a solution described by Dave Tang in this blog post.

pyconserve takes as input the following:

1. A BED file containing the intervals of interest
2. Phastcons or PhyloP bedGraph files (one per chromosome)

It then uses [pybedtools] to perform genomic intersections between the input BED file and bedGraph files. As each chromosome is queried independently, this can be run using parallelization to speed up runtime.

Note: this software is in beta and may contain bugs.

1. Install bedtools

2. Clone the pyconserve GitHub repo

    git clone https://www.github.com/kcha/pyconserve.git
    cd pyconserve
   

3. Install the package

    python setup.py install
   

The following steps describe how to download conservation tracks and convert them to bedGraph format using the UCSC tools wigToBigWig and bigWigToBedGraph. These steps are adapted from this blog post.

1. Download Phastcons or PhyloP wiggle files (one chromosome per file) from UCSC

2. Convert wiggle to bigWig using wigToBigWig.

    wigToBigWig chr1.phastCons100way.wigFix.gz hg19.chrom_sizes.txt \
        chr1.phastCons100way.bigWig
   

   This step requires a file containing the chromosome sizes of your species.

3. Convert bigWig to bedGraph using bigWigToBedGraph

    bigWigToBedGraph chr1.phastCons100way.bigWig chr1.phastCons100way.bedGraph
   

4. (optional) To save space, compress the bedGraph file and remove the wiggle and bigWig files

    rm -v chr1.phastCons100way.wigFix.gz chr1.phastCons100way.bigWig
    gzip -vf chr1.phastCons100way.bedGraph
   

Given a BED file and a set of bedGraph conservation files, pyconserve can be run as follows:

pyconserve a.bed chr*.phastCons100way.bedGraph.gz > a_conservation.bed

This will perform bedtools intersect and save the output to a_conservation.bed.

To aggregate these results by computing the mean conservation score for each interval in the BED file, use the command summarize_conserve:

summarize_conserve a_conservation.bed > summarized.bed

Alternatively, the above two commands can be chained together as follows:

pyconserve a.bed chr*.phastCons100way.bedGraph.gz | \
    summarize_conserve - > summarized.bed

This software is inspired by previous work from others:

• Extracting sequence conservation: https://davetang.org/muse/2012/08/07/sequence-conservation-in-vertebrates/
• Multiprocessing using pybedtools: https://daler.github.io/pybedtools/3-brief-examples.html#example-3-count-reads-in-introns-and-exons-in-parallel