Transposable Elements Pipeline Analysis

Introduction

This repository contains a collection of pipelines and scripts to analyze RNA-seq and ChIP-seq data taking into account multi-mapping reads for accurate transposable elements quantification.

RNA

TEtranscript toolkit is used for multi-mappers-awar family-level quantifications and for differential analyses.

TElocal is used for multi-mappers-awar loci-level quantifications.

Histone modifcations

SmartMap is used for multi-mappers-awar quantifications.

Getting started

The first step is to download this repository to your computer using the following commands:

$ git clone [email protected]:broadinstitute/broad-epi-repeats-analysis.git
$ cd broad-epi-repeats-analysis

Genome, genes, and repeats annotations

The script in src/bash/create-mus-musculus-annotations-mm10.bash will download and create all the necessary annotations. The annotations will be place in a new folder mm10 with three sub-folders called genome, genes, and repeats.

Genome index

Using the genome FASTA file insided mm10/genome you can build the index for STAR and bowtie2 using the star-build-index wdl or the bowtie2-build-index wdl respectively. Both are located in the workflows folder.

Quantifications

workflow/align-quantify-repeats.wdl workflow will align the FASTQs to the genome using STAR and compute genes and TEs quantifications using TElocal and TEcount. Four count files will be reported:

• family-level-unique-counts
• family-level-multimappers-counts
• loci-level-unique-counts
• loci-level-multimappers-counts.

bowtie2-repeats-quantification workflow will align the FASTQs to the genome using bowtie2 and a weighted bedgraph accounting for multi-mappers will be generated. gs://encode-pipeline-genome-data/mm10/bowtie2_index/ENCFF309GLL.tar.gz gs://encode-pipeline-genome-data/mm10/mm10_no_alt.chrom.sizes.tsv

smartmapprep outputp is a bed.gz file that ends with ("coord_scores.bed.gz") and a folder called splits.

TODO:

• Report unique-mappers track for ChIP
• Create tracks for RNA using unique-mappers and apportioning multi-mappers. It will not be the same counts as in the count files but it will give us an idea.
• Annotate each loci with the percentage of gained counts when including multi-mappers. If unique counts were 3 and with multi-mappers is 100 it is suspicious compared to something going from 50 to 100.