nova: de novo insertion simulator

Overview

nova simulates de novo variant insertions in long-read sequencing data.

The simulator can generate random, simple repeat, and predefined insertion sequences and insert them into reads at ultra-low frequencies typical of de novo structural variants.

Installation

conda (recommended)

We recommend using mamba or conda for (almost) one-step installation.

# we recommend using mamba for faster environment resolution
mamba env create -f environment.yml --subdir osx-64
conda activate nova
conda config --env --set subdir osx-64 

uv pip install -e. # switch to conda later, but uv recognizes and respects activate conda env

Standalone install

nova can be installed using uv or pip. We recommend installing it into a venv managed by uv, like so:

uv venv nova-venv
source nova-venv/bin/activate
uv pip install -e. # atm this is dev only, switch to nova later if up on pypi

The requirements for nova are detailed in both setup.py and requirements.txt. However, there are several dependencies that aren't available via PyPi, hence why we recommend conda/mamba.

If you want to avoid that path entirely, you'll need to install additional dependencies to use the snakemake pipeline:

bedtools==2.31.1
samtools==1.22
bcftools==1.22
minimap2==2.30
sniffles==2.6.1
truvari==5.3.0
snakemake-minimal==9.6.2

Quick Start

Read simulation

nova has one main program (nova simulate), which samples reads from an input BAM file and modifies them with insertion variants according to a configuration JSON file.

nova simulate input.bam config.json \
--random-seed 783 \
-o output_dir \
-p nova

For more options and info, read the quick-start simulate and configuration docs.

Analysis

nova is a variant simulator and doesn't perform variant calling and simulation analysis on its own. There are a number of accompanying scripts that you can use to assess your simulations, described in the analysis docs and run by default in the analysis pipeline.

Analysis pipeline

We recommend using the environment.yml file to install all requirements to run the full pipeline.

1. Prepare your BAM file and variant configuration JSON file specifying desired variants
2. Prepare the config file (template in examples/snakemake_config.yml).
   • Note: You will need to self-supply the path to the reference genome and repeat file (for sniffles2).
3. Dry run the pipeline, specifying the path to your edited config file:

snakemake --use-conda --configfile {edited_config.yml} -np

4. Once the dry run is working, run the pipeline with resource-appropriate core allocation:

snakemake --use-conda --configfile {edited_config.yml} --cores {cores}

License

(tbd?)