1. Introduction
2. Build Resources
   2.1 Bowtie2 Indices
   2.2 Image from Dockerfile
3. Run HERVx pipeline
   3.1 Using Singularity
   3.2 Using Docker
   3.3 Using WDL and Cromwell
4. TLDR
5. References

Human Endogenous Retrovirus expression pipeline, as known as HERVx, is a containerized pipeline to characterize retrotranscriptome. Quantifying HERV expression is difficult due to their repetitive nature and the high degree of sequence similarity shared among subfamilies— leading to an inherit level of uncertainty during fragment assignment.

HERVx calculates Human Endogenous Retrovirus (HERV) expression in paired-end RNA-sequencing data. The pipeline runs cutadapt¹ to remove adapter sequences and to perform quality-trimming, bowtie2² to align reads against the Human reference genome (hg38), SAMtools³ to convert from SAM to BAM format and to sort reads by name, and Telescope⁴ to characterize Human Endogenous Retrovirus (HERV) expression.

Telescope is a computational method that provides accurate estimation of transposable element expression. It directly addresses uncertainty in fragment assignment by reassigning ambiguously mapped fragments to the most probable source transcript as determined within a Bayesian statistical model.

The Dockerfile will build cutadapt, bowtie2, SAMtools & HTSlib, and Telescope from scratch along with a few other tools. Small reference files are located in /opt2/refs/ in the container's filesystem.

Reference files, resources, and indices are bundled within the container's filesystem.

Currently, the following files are located in /opt2/refs/:

• trimmonatic_TruSeqv3_adapters.fa
• HERV_rmsk.hg38.v2.genes.gtf
• HERV_rmsk.hg38.v2.transcripts.gtf
• L1Base.hg38.v1.transcripts.gtf
• retro.hg38.v1.transcripts.gtf

Bowtie2 indices for hg38 are bundled in the container's filesystem in /opt2/bowtie2/. Other indices can be provided by mounting the host filesystem to this PATH (overrides current hg38 indices).

# Get UCSC hg38 genome
wget http:https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/hg38.fa.gz
zcat hg38.fa.gz > hg38.fa

# Build the indices
module load singularity
SINGULARITY_CACHEDIR=$PWD singularity pull -F docker:https://nciccbr/ccbr_telescope
singularity exec -B $PWD:$PWD ccbr_telescope_latest.sif bowtie2-build hg38.fa hg38

In the example below, change skchronicles with your DockerHub username.

# See listing of images on computer
docker image ls

# Build
docker build --tag=ccbr_telescope:v0.0.1 .

# Updating tag(s) before pushing to DockerHub
docker tag ccbr_telescope:v0.0.1 skchronicles/ccbr_telescope:v0.0.1
docker tag ccbr_telescope:v0.0.1 skchronicles/ccbr_telescope        # latest
docker tag ccbr_telescope:v0.0.1 nciccbr/ccbr_telescope:v0.0.1
docker tag ccbr_telescope:v0.0.1 nciccbr/ccbr_telescope             # latest

# Check out new tag(s)
docker image ls

# Peak around the container: verify things run correctly
docker run -ti ccbr_telescope:v0.0.1 /bin/bash

# Push new tagged image to DockerHub
docker push skchronicles/ccbr_telescope:v0.0.1
docker push skchronicles/ccbr_telescope:latest
docker push nciccbr/ccbr_telescope:v0.0.1
docker push nciccbr/ccbr_telescope:latest

module load singularity
# Pull from DockerHub
SINGULARITY_CACHEDIR=$PWD singularity pull -F docker:https://nciccbr/ccbr_telescope
# Display usage and help information
singularity exec -B $PWD:$PWD ccbr_telescope_latest.sif HERVx -h
# Run HERVx pipeline
singularity exec -B $PWD:$PWD ccbr_telescope_latest.sif HERVx -r1 tests/small_S25_1.fastq -r2 tests/small_S25_2.fastq -o ERV_hg38

# Assumes docker in $PATH
docker run -v $PWD:/data2 nciccbr/ccbr_telescope:latest HERVx -r1 tests/small_S25.R1.fastq.gz -r2 tests/small_S25.R2.fastq.gz -o ERV_hg38

# hervx is configured to use different cromwell execution backends: local or slurm
# view the help page for more information
./hervx --help

# @local: uses local singularity cromwell backend
# The local EXECUTOR will run serially on compute
# instance. This is useful for testing, debugging,
# or when a users does not have access to a high
# performance computing environment.
./hervx local -r1 tests/small_S25.R1.fastq.gz -r2 tests/small_S25.R2.fastq.gz --outdir ERV_hg38

# @slurm: uses slurm and singularity cromwell backend
# The slurm EXECUTOR will submit jobs to the cluster.
# It is recommended running hervx in this mode.
./hervx slurm -r1 tests/small_S25.R1.fastq.gz -r2 tests/small_S25.R2.fastq.gz --outdir ERV_hg38

Reference files are located in /opt2/ of the container filesystem.
Dockerfile to build this image is located in /opt2/Dockerfile.
Pull latest image from DockerHub
Usage
 singularity exec docker:https://nciccbr/ccbr_telescope HERVx -h
 docker run nciccbr/ccbr_telescope:latest HERVx -h

<sup>1. Martin, M. (2011). "Cutadapt removes adapter sequences from high-throughput sequencing reads." EMBnet 17(1): 10-12.
2. Langmead, B. and S. L. Salzberg (2012). "Fast gapped-read alignment with Bowtie 2." Nat Methods 9(4): 357-359.
3. Li, H., et al. (2009). "The Sequence Alignment/Map format and SAMtools." Bioinformatics 25(16): 2078-2079.
4. Bendall, M. L., et al. (2019). "Telescope: Characterization of the retrotranscriptome by accurate estimation of transposable element expression." PLOS Computational Biology 15(9): e1006453.</sup>

Back to Top