NOTE: This repository contains the experiments described in our reviesed manuscript completed in the end of May 2021. In case there are any problems running the experiments, please contact us by e-mail.
PanVC is a set of tools to be used as part of a variant calling workflow that uses short reads as its input. The reads are aligned to an index generated from a multiple sequence alignment. A suitable index may be built from founder sequences.
Running a variant calling workflow that utilizes PanVC consists of the following phases:
- Generating founder sequences from known variants
- Indexing the founder sequences
- Running the read alignment and variant calling workflow
Our software consists of the following components:
- vcf2multialign for generating founder sequences and predicted sequences from known variants
- PanVC includes tools for various phases of the variant calling workflow, including indexing the known sequences, aligning short reads to the index and generating an ad-hoc reference sequence for re-aligning the reads
- PanVC sample workflow for running a variant calling workflow that utilizes PanVC
All necessary components for running the experiments from our provided inputs are installed by running Snakemake, which in turn invokes Conda. To install the components to a predefined location, e.g. in conda-env in the root of the cloned repository, please run the following commands. (By default, i.e. when --conda-prefix is not given, the Conda environment is placed in a hidden .snakemake directory in the working directory when Snakemake is run. This may be less convenient for running multiple experiments.)
Please note, however, that prebuilt binaries for some of the software are only available for Linux on x86-64.
- Clone the repository with
git clone --recursive https://github.com/algbio/panvc-founders.git cd panvc-founders- Prepare the Conda environments with the following command:
snakemake --cores 1 --printshellcmds --use-conda --conda-prefix ./conda-env conda_environment conda_environment_gatk conda_environment_edlib conda_environment_happy conda_environment_vcfcat
On a high level, the experiments are run as follows:
- Decide which experiments to run and possibly modify the configuration files in the subdirectory of the corresponding experiment.
- Download and unarchive the reads provided with the experiments
- Do one of the following:
- Download and unarchive the pre-generated indices provided with the experiments in question and call variants by running Snakemake
- Download and decompress the founder sequences provided with the experiments in question, generate the index and call variants by running Snakemake
- Process the variant calling results to e.g. calculate edit distances from a specific sequence or calculate statistics from the called variants.
We provide scripts to automatize these steps.
Each experiment involves aligning different sets of reads to different indices. Inputs for the artificial mutation experiment are the smallest and therefore parts of the experiment should be the fastest to run.
Please see README.md under experiments-with-artificial-mutations.
Please see README.md under e-coli.
Please see README.md under take-one-out-experiment-with-human-chr-1.
Please see README.md under scalability-experiment.