Custom reference data

The pipeline uses reference data at various stages, such as for alignment, calculating TSS enrichments, and other QC scores. If you're using a common genome assembly, these resources are pre-built and can be easily downloaded using refgenie pull, as described in the setup instructions. If the resources are not available, you'll have to build them. This document outlines how we created the reference data, so you can recreate it if you need to. The easiest way to do this is use refgenie build. All you need to do is:

1: Build the fasta asset

You need a FASTA file for your genome. You can insert this file into refgenie like this:

refgenie build -g GENOME -a fasta --files fasta=/path/to/file.fa

2. Build the bowtie2_index

To build a bowtie2_index and have it managed by refgenie you'll, of course, need bowtie2 already installed. You will also need the requisite FASTA file, which you just added in step 1.

refgenie build -g GENOME -a bowtie2_index

3: Build the ensembl_gtf asset

The ensembl_gtf asset includes several related assets (e.g. pause index gene bodies and TSS's) the pipeline will employ. To build an ensembl_gtf asset, you need an Ensembl GTF file (or equivalent) for your genome. You can have refgenie build and manage this file as follows:

refgenie build -g GENOME -a ensembl-gtf --files ensembl_gtf=/path/to/Homo_sapiens.GRCh38.97.gtf.gz

4: Build the refgene_anno asset

The refgene_anno asset actually includes several related assets that we'll need (e.g. TSS and premature mRNA annotations). To build these, for example for hg38, you will need to download a refGene annotation. Build it for a any genome like so:

refgenie build -g GENOME -a refgene_anno --files refgene=/path/to/refGene.txt.gz

5: Build the feat_annotation asset

The feat_annotation asset includes feature annotations used to calculate the FRiF and cFRiF. Refgenie can automatically build this after you have the above assets installed:

refgenie build -g GENOME -a feat_annotation

That's it! These assets will be automatically detected by PEPPRO if you build them like this with refgenie.

Create a custom feature annotation file

The pipeline will calculate the fraction (and proportion) of reads in genomic features using the feat_annotation asset, but you can also construct this file yourself.

This annotation file is really just a modified BED file, with the chromosomal coordinates and type of feature included. For example, the downloadable hg38_annotations.bed.gz file looks like so:

chr1    28200   30001   Promoter    .   *
chr1    198800  200201  Promoter    .   *
chr1    778000  780001  Promoter    .   *
chr1    817400  817601  Promoter    .   *
chr1    826200  828801  Promoter    .   *
chr1    904200  905201  Promoter    .   *
chr1    923800  924601  Promoter    .   *
chr1    925000  925601  Promoter    .   *
chr1    941800  942201  Promoter    .   *
chr1    958400  961401  Promoter    .   *

Just like a standard BED file, the first three fields are:
1. chrom - the name of the chromosome
2. chromStart - the starting position of the feature
3. chromEnd - the ending position of the feature

Column four is the name column, in our case the name of our feature of interest. The fifth column is the score, which would determine how darkly an item would be displayed in a genome browser if you chose to set that or if the information in your file of interest has ascribed a score to the features. The final, sixth, column is the strand column.

After creating your BED file, you can point the pipeline to it using the --anno-name option followed with the path to your file. The pipeline will then use that file to determine the fractions of reads that cover those features.