Introduction

The following repository contains the scripts used to process that data as well as generate the figure in "Leveraging histone modifications to improve genome annotations" - Mendieta et al 2021.

The corresponding directories are broken down by script type, with the snakemake directory being further broken down by analysis type.

Annotating Genomes Using Chromatin Modification Data (Script)

Below is a detailed example of how to utilize the annotation script implemented in Mendieta et al. Detailed infomration about pre-processing ChIP-seq data to utilize this script is included in the script update_annotation_by_chromatin_mods.snake. But paired down, the script requires a set a genome annotations in bed format (genes and lncRNAs). A set of peaks from ChIP-seq peak calling focused on modifications corresponding to elongation which should occur throughout the gene body (H3K36me3, H3K4me1), as well as initiation modifications which correlate with the promoter of a gene (H3K4me3, H3K56ac). At least one modification of each type is required. Additionally stranded RNA-seq data is required to ensure that non-expressed gene features overlapping our modifications will be ignored. Finally a bigwig file corresponding to one of the elongation mods is helpful, as it allows the script to identify large genes while accounting for potential mappability issues, as well as lack of upstream signal from the elongation mod.

Basic Script Command:

python scripts/Update_annotation.py -broad {input.broad} \
-narrow {input.narrow_peaks} -annotation {input.gene_annotation} \
-bw {input.H3K36me3_bw} \
-RNA {input.rna_file} -lncRNA {input.gene_annotation} -o {output}

A list of test/example files can be found in this github repo under the file example_data.

The output of this script is a list of annotations - with their potential annotation corrected class append as original.annot_ANNOTATION_CLASS. Potential novel annotations are awarded the name Possible_novel_gene.

Note: This script is implemented in python 3.6, and requires pybedtools, and pysam. Both of which can be easily installed using either homebrew or Anaconda3. Additionally since many genomes which this analysis can be run on are fairly large, they will need to be broken up on a chromosome by chromosome basis. This ensures that pysam is able to read the full RNA-seq BAM file without overflowing.

Annotating Genomes Using Chromating Modification Data (Pipeline)

The implemenation of this method is further detailed in the snakemake script located snakemake_scripts/12.Annotate_other_species. Here you will find an example implementation in soybean (glycine max). Structure of the inputs can be gleaned from the snakemake config yaml file, as well as a sample file corresponding to the features mentioned above. Note that tissues need to be matched in the sample file. This snakemake analysis was done in version of 5.7.1. Since then aggregate command behavior is slighly modified, so be aware that running this analysis on more updated version of snakemake may result in unexpected behaviors.