Skip to main content
Version: 3.18

Getting Started

Nirvana is written in C# using .NET Core (an amazing runtime environment that currently runs on Windows, Linux, Mac OS X, and in Docker images). Once .NET Core has been downloaded, all you need to do is grab the source, compile it, and grab the data files.

tip

Nirvana currently uses .NET Core 3.1 or later. Please make sure that you have the most current runtime from the .NET Core downloads page.

Quick Start

If you want to get started right away, we've created a script that downloads Nirvana, compiles it, and starts annotating a test file:

curl -O https://illumina.github.io/NirvanaDocumentation/files/TestNirvana.sh
bash ./TestNirvana.sh

We have verified that this script works on Windows (using Git Bash or WSL), Linux, and Mac OS X.

Getting Nirvana

Compile from Source

The following will grab the latest version of Nirvana from GitHub and compile it using the .NET Core compiler:

git clone https://github.com/Illumina/Nirvana.git
cd Nirvana
dotnet build -c Release

GitHub Release Notes

Alternatively, you can grab the latest binaries from our GitHub Releases page:

mkdir -p Nirvana/Data
cd Nirvana
unzip Nirvana-3.16.1-dotnet-3.1.0.zip

Docker

You can find us on Docker Hub under annotation/nirvana:

caution

We think Docker is fantastic. However, because our data files are usually accessed through a Docker volume, there is a noticeable performance penalty when running Nirvana in Docker.

mkdir -p Nirvana/Data
cd Nirvana
docker pull annotation/nirvana:3.14

For Docker, we have special instructions for running the Downloader:

sudo docker run --rm -it -v Data:/scratch annotation/nirvana:3.14 dotnet \
     /opt/nirvana/Downloader.dll --ga GRCh37 -o /scratch

Similarly, we have special instructions for running Nirvana (Here's a toy VCF in case you need it):

sudo docker run --rm -it -v Data:/scratch annotation/nirvana:3.14 dotnet \
     /opt/nirvana/Nirvana.dll -c /scratch/Cache/GRCh37/Both \
     -r /scratch/References/Homo_sapiens.GRCh37.Nirvana.dat \
     --sd /scratch/SupplementaryAnnotation/GRCh37 \
     -i /scratch/HiSeq.10000.vcf.gz -o /scratch/HiSeq

Downloading the data files

To download the latest data sources (or update the ones that you already have), use the following command to automate the download from S3:

dotnet bin/Release/netcoreapp3.1/Downloader.dll \
     --ga GRCh37 \
     -o Data
  • the --ga argument specifies the genome assembly which can be GRCh37, GRCh38, or both.
  • the -o argument specifies the output directory
Glitches in the Matrix

Every once in a while, the download process does not go smoothly. Perhaps the internet connection cut out or you ran out of disk space. The Downloader attempts to detect these situations by checking the file sizes at the very end. If you see that a file was marked truncated, try fixing the root cause and running the downloader again.

tip

From time to time, you can re-run the Downloader to get the latest annotation files. It will only download the files that changed.

Download a test VCF file

Here's a toy VCF file you can play around with:

curl -O https://illumina.github.io/NirvanaDocumentation/files/HiSeq.10000.vcf.gz

Running Nirvana

Once you have downloaded the data sets, use the following command to annotate your VCF:

dotnet bin/Release/netcoreapp3.1/Nirvana.dll \
     -c Data/Cache/GRCh37/Both \
     --sd Data/SupplementaryAnnotation/GRCh37 \
     -r Data/References/Homo_sapiens.GRCh37.Nirvana.dat \
     -i HiSeq.10000.vcf.gz \
     -o HiSeq.10000
  • the -c argument specifies the cache prefix
  • the --sd argument specifies the supplementary annotation directory
  • the -r argument specifies the compressed reference path
  • the -i argument specifies the input VCF path
  • the -o argument specifies the output filename prefix

When running Nirvana, performance metrics are shown as it evaluates each chromosome in the input VCF file:

---------------------------------------------------------------------------
Nirvana                                             (c) 2021 Illumina, Inc.
Stromberg, Roy, Lajugie, Jiang, Li, and Kang                         3.16.1
---------------------------------------------------------------------------

Initialization                                         Time     Positions/s
---------------------------------------------------------------------------
Cache                                               00:00:01.2
SA Position Scan                                    00:00:00.1       55,270

Reference                                Preload    Annotation   Variants/s
---------------------------------------------------------------------------
chr1                                    00:00:00.1  00:00:01.5        6,323

Summary                                                Time         Percent
---------------------------------------------------------------------------
Initialization                                      00:00:01.3       23.9 %
Preload                                             00:00:00.1        2.9 %
Annotation                                          00:00:01.5       27.2 %

Peak memory usage: 1.434 GB
Time: 00:00:05.2

The output will be a JSON file called HiSeq.10000.json.gz. Here's the full JSON file.