proseg-workflow

Basic nextflow pipeline to run Proseg on a single sample. After Proseg segmentation, the output is converted into Xenium Explorer compatible format

NOTE: This pipeline currently uses proseg v1 which is not suitable for highly multiplex datasets (i.e. Xenium Prime 5k panel). See this docker container for proseg v2 which is designed for highly multiplex data.

Configuration

Pipeline parameters can be specified in nextflow.config

Key parameters include:

• platform - xenium, cosmx, or merfish
• outputDir - where to store the output files
• inputDir - directory with the input file(s)
• transcriptsFile - filename in inputDir with transcript info. For Xenium this is transcripts.parquet
• enforce_connectivity - prevent cells from having disconnected voxels
• ignore_z_coord - run in 2D mode
• sampleID - name of the sample for Xenium Ranger output

CosMx

Right now proseg-workflow only supports running Proseg on CosMx data with the --cosmx flag. If there is interest in supporting --cosmx-micron and FOV stitching, please create an issue.

Install

Make sure Nextflow and Proseg are installed.

Docker

A docker container with Proseg is provided easy deployment

To run the pipeline with docker:

nextflow run main.nf -with-docker [docker env name]

Execution

This workflow can be run on Cirro

On Cirro the workflow is launched with 16 CPUs and 64GB of memory on AWS. If the initial run crashes, nextflow will try to relaunch 3 times with increasing levels of memory (64GB, 128GB, 192GB).