SVTeaser

SV simulation for rapid benchmarking

Previous Work

Hackathon Schedule

Presentation

Goals

Make a tool that (A) performs SV and read simulation to create inputs for benchmarking an SV caller (B) creates an evaluation/reporting of the SV caller's performance. Users supply SVTeaser with a reference sequence file (.fasta) and, optionally, a set of SVs (.vcf). SVTeaser outputs assorted statistical metrics across a range of read lengths and depths. SVTeaser achieves rapid assessment by downsampling the full reference to a subset of numerous 10kb samples to which it adds SVs.

Overview Diagram

Working Notes/Documentation

Here

File structure diagram

Define paths, variable names, etc

Installation

• Build the SVTeaser pip install-able tarball

• Download and install SURVIVOR

• Put the SURVIVOR executable into your environment's PATH

• The three steps of this are handled by bash install.sh

• Install vcftools

• Ensure vcftools (e.g. vcf-sort) is in your environment's PATH

• Put ART read simulator executable into your environment's PATH

• Install truvari

Quick Start

usage: svteaser [-h] CMD ...

SVTeaser v0.0.1 - SV simulation for rapid benchmarking

    CMDs:
        sim_sv          Simulate SVs
        surv_sim        Simulate SVs with SURVIVOR
        surv_vcf_fmt    Correct a SURVIVOR simSV vcf
        sim_reads       Run read simulators

positional arguments:
  CMD         Command to execute
  OPTIONS     Options to pass to the command

optional arguments:
  -h, --help  show this help message and exit

Workflow:

• Create a SVTeaser working directory (output.svt) by simulating SVs over a reference

• svteaser surv_sim reference.fasta workdir

2. in progress Simulate reads over the altered reference and place them in the output.svt directory

• svteaser sim_reads workdir.svt

3. Call SVs over the reads (output.svt/read1.fastq output.svt/read2.fastq) with your favorite SV caller
4. Run truvari bench with the --base output.svt/simulated.sv.vcf.gz and --comp your_calls.vcf.gz
5. Open the notebooks/SVTeaser.ipynb and point to your output.svt directory

See test/workflow_test.sh for an example

Component Details

SV Simulator

Two methods for SV simulation are supported in SVTeaser - (done) simulation of SV with SURVIVOR and (in progress) simulation of SVs from VCFs.

Running simulation in either mode results in an output directory of the following structure -

$ svteaser surv_sim reference.fasta workdir
$ ll -h workdir
total 2.3M
drwxr-xr-x  2 user hardware 4.0K Oct 12 15:38 ./
drwxr-xr-x 13 user hardware 4.0K Oct 12 15:38 ../
-rw-r--r--  1 user hardware 1.1M Oct 12 15:38 svteaser.altered.fa # <---- Multi-FASTA with all altered region sequences
-rw-r--r--  1 user hardware 980K Oct 12 15:38 svteaser.ref.fa     # <---- Multi-FASTA with all unaltered region sequences
-rw-r--r--  1 user hardware 228K Oct 12 15:38 svteaser.sim.vcf    # <---- Combined VCF with variants from each region
-rw-r--r--  1 user hardware  34K Oct 12 15:38 svteaser.sim.vcf.gz
-rw-r--r--  1 user hardware  121 Oct 12 15:38 svteaser.sim.vcf.gz.tbi