Exogene

---

Exogene

A workflow for detecting viral integrations from both short read and long read sequencing data.

---

usage

docker pull zstephens/exogene:v15

docker run -it -v ${HOME}:${HOME} zstephens/exogene:v15

Create human + viral reference sequence:

(from inside the container)

/home/init_ref.sh \
    -i /path/to/hg38.fa \
    -o /path/to/hg38_plus_viral.fa

Using custom viral references:

If you wish to use viral reference sequences different than what Exogene uses by default, you can use the -v input option to specify a fasta file of viral genomes. Exogene expects that contigs are named in the following format:

>accession_id full_name (space delimited)

For example: >NC_009334 Human herpesvirus 4, complete genome.

Additionally, the viral fasta should be indexed using bwa index

Running Exogene-SR (with BAM input)

/home/Exogene-SR.sh \
    -b input.bam \
    -r hg38_plus_viral.fa \
    -o outDir/

If custom viral sequences were used, the -v input option will be required.

Running Exogene-SR (with FQ input)

/home/Exogene-SR.sh \
    -f1 read1.fq.gz \
    -f2 read2.fq.gz \
    -r hg38_plus_viral.fa \
    -o outDir/

Input FASTQ files must be gzipped. Currently Exogene-SR does not support single-end reads. If custom viral sequences were used, the -v input option will be required.

Running Exogene-LR (with FASTQ input, e.g. PacBio HiFi reads)

/home/Exogene-LR.sh \
    -f input.fq.gz \
    -r hg38_plus_viral.fa \
    -m hifi \
    -o outDir/

Running Exogene-LR (with FASTA input, e.g. PacBio CLR reads)

/home/Exogene-LR.sh \
    -f input.fa.gz \
    -r hg38_plus_viral.fa \
    -m clr \
    -o outDir/

Running Exogene-LR (with BAM input)

/home/Exogene-LR.sh \
    -b input.bam \
    -r hg38_plus_viral.fa \
    -m [hifi/clr] \
    -o outDir/

Intersecting Exogene-SR and Exogene-LR results:

python /home/combine_reports.py \
    -s Viral_Reads_Report.tsv \
    -l Viral_Junctions_LongReads.tsv \
    -o combined_report_outDir/ \
    -ms minimum_number_of_softclipped_reads_per_site [1] \
    -md minimum_number_of_discordant_pairs_per_site [5]

Either -s or -l must be specified (or both, for a combined report). Viral_Reads_Report.tsv is created in the output directory of Exogene-SR, Viral_Junctions_LongReads.tsv is created in the output directory of Exogene-LR.

Test Data:

The Docker container contains a small quantity of test data which can be processed as follows:

/home/Exogene-SR.sh \
    -f1 /home/test_data/SRR3104446_1.fq.gz \
    -f2 /home/test_data/SRR3104446_2.fq.gz \
    -r /path/to/hg38_and_viral.fa \
    -o /path/to/out_SR/

/home/Exogene-LR.sh \
    -f /home/test_data/a1el_ccs.fq.gz \
    -r /path/to/hg38_and_viral.fa \
    -m hifi \
    -o /path/to/out_LR/

For the included hg38+viral reference, the bwa/pbmm2 alignment steps require ~32GB of memory.