sequintools

sequintools is a suite of tools for manipulating and reporting on Next Generation Sequencing data that has Sequins added to the sample.

Installation

If you're a Rust programmer, sequintools can be installed with cargo.

cargo install sequintools

Alternatively, one can use cargo binstall to install a sequintools binary directly from GitHub:

cargo binstall sequintools

Note that sequintools has a dependency on cmake and clang, and these will need to be installed on your device before you can install sequintools. On a Linux device you would usually install these using apt, e.g.

sudo apt update && sudo apt install -y cmake clang

If you're not using a Linux device then you'll need to research how to install the dependencies on your device. Alternatively you can use the docker build which will already have the dependencies required installed.

Docker

We also provide sequintools as a Docker container.

This command line below is used to fetch its latest version.

docker pull ghcr.io/sequinsbio/sequintools:latest

As an example, you could run the calibration step as below, make sure the example folder is under your current path.

docker run --rm -v $PWD:/data ghcr.io/sequinsbio/sequintools \
    sequintools calibrate \
    -b /data/example/resources/sequin_regions.chrQ_mirror.bed \
    -f 40 \
    -o /data/calibrated.bam \
    --write-index \
    /data/example/example.bam

Building

sequintools is written in Rust, so you'll need to grab a Rust installation in order to compile it.

To build sequintools:

git clone https://github.com/sequinsbio/sequintools
cd sequintools
cargo build --release
sudo cp target/release/sequintools /usr/local/bin

Usage

To get going, we provide some example data in the example directory. This is simulated data that will allow you to quickly run commands from sequintools and familiarise yourself with the input file requirements.

calibrate

calibrate downsamples Sequins reads so that they more closely match the sample data. The more closely the Sequins data resembles the sample data it controls for, the better the control will be.

The simplest form of calibration is to adjust the mean coverage for all Sequins regions to a flat target coverage. In this example, the mean coverage for each Sequins molecule in the data set is set to 40X.

sequintools calibrate \
    -b example/resources/sequin_regions.chrQ_mirror.bed \
    -f 40 \
    -o calibrated.bam \
    --write-index \
    example/example.bam

Alternatively, you can use the sample data in the same BAM file to adjust the Sequins coverage to more closely represent the coverage of the controlled region. This method uses the mean depth of the region in the sample data that corresponds to each Sequins molecule rather than a fixed coverage for all Sequins regions. To do this, you need to provide two BED files: one containing the Sequins regions on the decoy chromosome and the other containing the reference genome locations they control for. The name column (4th column) must match between the two BED files.

sequintools calibrate \
    -b example/resources/sequin_regions.chrQ_mirror.bed \
    -H example/resources/sequin_regions.hg38.bed \
    -o calibrated.bam \
    --write-index \
    example/example.bam

Make the same adjustments as the previous command, but exclude the sample data so that the output BAM only has calibrated Sequins data (this is much faster than the previous command).

sequintools calibrate \
    -b example/resources/sequin_regions.chrQ_mirror.bed \
    -H example/resources/sequin_regions.hg38.bed \
    -o calibrated.bam \
    --write-index \
    --exclude-uncalibrated-reads \
    example/example.bam

Note

In samples with high mitochondrial DNA copy number, native chrM coverage may exceed the coverage of the corresponding mitochondrial Sequins (SG_000000038–SG_000000041) for the WGS Core Control Set product. Because calibrate can only downsample Sequins reads, sample-matched calibration will stop if the required target coverage is higher than the available Sequins coverage.

This does not necessarily indicate poor mitochondrial Sequins recovery or mapping. To calibrate the remaining regions, remove the mitochondrial Sequins from the decoy BED file. If required, mitochondrial Sequins reads can be extracted before calibration and merged back into the calibrated BAM or CRAM afterwards.

The same error for a nuclear region is more likely to indicate an input or workflow issue, as nuclear Sequins coverage is expected to exceed that of the corresponding native region.

bedcov

The bedcov command collects statistics from a BAM file for regions in a BED file and writes them as a CSV. Statistics include the minimum and maximum coverage per base, mean coverage, standard deviation of coverage and the coefficient of variance.

sequintools bedcov \
    example/resources/sequin_regions.hg38.bed \
    example/example.bam