Running Batch Jobs on GCP via dsub

What problem does it solve?

This Skill helps you execute compute-heavy viral NGS tasks on Google Cloud in a reliable, scalable way when running locally is too slow, too memory-intensive, or requires specialized Docker images.

Core Features & Use Cases

• Parallel batch execution with dsub: Runs many independent jobs from a single task TSV to maximize throughput.
• Dockerized tool execution: Uses a specified container image so tool versions and dependencies stay consistent.
• GCS-first data flow: Downloads declared inputs from GCS and uploads outputs/logs back to GCS for auditing and debugging.
• Common viral analysis scenarios: Useful for memory-heavy steps like VADR, BLAST, and genome assembly, especially when processing dozens of sequences or assemblies.

Quick Start

Run dsub with the GCS-based provider settings, a Docker image, a machine type, a script, and a tasks TSV so each TSV row launches an independent containerized job and writes logs to your GCS logging path.