MinION pipeline

Data sync

Primary sequencing data lives on the Rhino FHCRC cluster at:

/fh/fast/bedford_t/zika-seq/data/

And locally on the Meristem drive at:

/Volumes/Meristem/data/

To sync Meristem to Rhino, run:

rsync -azP <username>@rhino.fhcrc.org:/fh/fast/bedford_t/zika-seq/data/ /Volumes/Meristem/data/

Replacing <username> with your username. This data/ directory is assumed to follow a particular schema.

Basecalling reads

Reads are basecalled using a command-line interface version of Metrichor (available upon request from ONT).

We’ve encountered issues with function recursion that causes nodejs to throw this error: maximum call stack size exceeded. We get around this by increasing the stack size for node. Note that -w specifies the workflow ID, here 1289 is the numeric code for 2D basecalling with barcoding for FLO-MIN106.

node --stack-size=65500 ./metrichor-cli -a <APIkey> -w 1289 -f - -i <path/to/raw_reads/> -o <path/to/basecalled_reads>

Bioinformatic pipeline

Here, we use the ZiBRA project bioinformatic pipeline at zibraproject/zika-pipeline. This pipeline is instantiated in the Docker image zibra/zibra. Data processing is done entirely using Docker.

Data volume

Create a named data volume that mirrors local data/ to data/ within container:

docker create --name zibra-data -v /Volumes/Meristem/data:/data zibra/zibra

This is to get data into the Docker container. Note that the path to local directory has to be an absolute path. Notes on data schema are here.

Samples volume

Create a named data volume that mirrors local sample metadata samples/ to samples/ within container:

docker create --name zibra-samples -v /Volumes/Meristem/samples:/samples zibra/zibra

This is to get sample metadata into the Docker container. Note that the path to local directory has to be an absolute path. Notes on metadata schema are here.

Build volume

Create a named data volume that mirrors local build/ to build/ within container:

docker create --name zibra-build -v /Volumes/Meristem/build:/build zibra/zibra

This is to get data out of the Docker container. Note that the path to local directory has to be an absolute path.

Start

Enter docker image:

docker run -t -i --volumes-from zibra-data --volumes-from zibra-samples --volumes-from zibra-build zibra/zibra /bin/bash

Run script:

python scripts/pipeline.py