APGAP and Seqera Platform — A Primer#

Arizona Pathogen Genomics Analytics Platform

If you’ve used APGAP for any length of time, you’ve probably noticed that the Pipelines section of a project hands off to something called Seqera Platform. This primer explains what Seqera is, why APGAP uses it, and what you can expect to see when working across the two systems.

This document is intended for technical staff and informed non-technical stakeholders. It assumes some familiarity with the idea of running computational pipelines — but no specific Nextflow or genomics tooling experience is required.

What Seqera Platform is#

Seqera Platform is a workflow orchestration and management product for running bioinformatics pipelines at scale. It was previously known as Nextflow Tower and was renamed to Seqera Platform in late 2023.

Seqera is built by the creators of Nextflow — the open-source workflow language that has become a de facto standard for reproducible bioinformatics analyses. Nextflow describes a pipeline as a graph of containerized steps and runs that graph across whatever execution environment you point it at: a single laptop, an HPC cluster, or thousands of cloud VMs. Seqera Platform sits on top of Nextflow and provides the missing pieces that turn it from a command-line tool into something an organization can actually operate at scale:

A web UI for launching pipelines without writing Nextflow command lines
Centralized monitoring, logging, and resource usage tracking for every run
Compute environment management — connecting Nextflow to AWS Batch, Google Batch, Kubernetes, Slurm, and similar targets
A workspace and organization model with role-based access, shared credentials, and pipeline catalogs
Reproducibility features (every run captures its exact configuration, container versions, and inputs)

Seqera supports the most widely used bioinformatics pipeline ecosystem available — the nf-core community library — alongside any custom Nextflow workflows your organization writes.

Why APGAP integrates with Seqera#

APGAP is fundamentally a data management platform: organize sequencing files, track metadata, control access, request datasets. Running the actual analyses — variant calling, assembly, lineage assignment, surveillance pipelines — is a separate problem with very different operational characteristics: long-running jobs, large compute clusters, container management, GPU scheduling, restart-on-failure logic, terabytes of intermediate files.

Building all of that into APGAP would have meant reinventing what Nextflow and Seqera already do well. Instead, APGAP delegates pipeline execution to Seqera and focuses on what it’s uniquely positioned to do: structured genomic data management for the public health context.

The integration follows a clean separation of concerns:

What APGAP handles	What Seqera handles
Sequence file storage and lifecycle	Pipeline execution
Metadata and source-type schemas	Compute environment management
Organizational hierarchy and RBAC	Workflow run history and logs
Analytical Dataset assembly and access requests	Container and dependency management
Audit trails for data access	Resource usage tracking
Notification system	Pipeline catalog

When you launch a pipeline from APGAP, APGAP packages your Analytical Dataset as input and tells Seqera to run a specific pipeline against it. Seqera does the actual work and writes results back to a GCS location that APGAP can read. APGAP then surfaces the result files in your project.

Per-project Seqera workspaces#

Seqera Platform uses a hierarchical model of organizations containing workspaces. A workspace is the unit of access control and resource scoping — pipelines, compute environments, credentials, and run history all live inside a workspace.

APGAP maps one Seqera workspace per APGAP project. When you create a project in APGAP, the platform automatically:

Triggers a Cloud Build pipeline that provisions a dedicated GCP environment (VPC, NAT, Vertex AI notebook instance, service accounts, IAM bindings)
Creates a Seqera workspace tied to that GCP environment
Configures the workspace with a compute environment pointing at your project’s GCP resources
Stores the workspace’s launchpad URL on the APGAP project record

This per-project model means every analysis project has isolated compute, isolated credentials, and isolated run history. Users assigned to a project’s Bioinformatics User role get access to that project’s Seqera workspace; users from other projects do not.

Provisioning typically takes a few minutes — somewhere in the 5 to 15 minute range, similar to how Lab provisioning works.

The Launchpad and pipeline catalog#

Inside each Seqera workspace is a Launchpad — Seqera’s name for the catalog of pre-configured pipelines that workspace members can run.

Each Launchpad entry combines three things into a one-click launch experience:

A workflow repository — a Git repository containing the Nextflow pipeline code (typically an nf-core pipeline like nf-core/viralrecon or nf-core/nascent, or a custom internal pipeline)
A set of default parameters — pre-filled values for the pipeline’s required inputs
A compute environment — the cloud or HPC target where the pipeline will actually run

Without the Launchpad, a researcher launching a Nextflow pipeline would need to specify the Git repo, version, container registry, compute target, and parameter overrides every time. The Launchpad bundles all of that into a named pipeline that can be launched in two clicks.

For APGAP projects, the Launchpad is typically pre-populated with the pipelines relevant to your source type — viral surveillance pipelines for clinical isolates, wastewater pipelines for environmental samples, and so on. The pipelines available to you depend on what your Platform Admin has configured for your project.

What you see in APGAP vs. inside Seqera#

The integration is designed so that most day-to-day work stays in APGAP, with Seqera taking over for the parts where it has a better UI or richer information.

You’ll do this in APGAP:

Upload sequence files and add metadata
Create Analytical Datasets that group files for analysis
Click Pipelines → Launch to start a run
See a summary of recent runs and their status
Browse and download result files when a run completes

You’ll typically jump to Seqera for:

Detailed run monitoring while a pipeline is executing
Process-level logs (each step of the pipeline has its own log)
Resource usage graphs (CPU, memory, disk, runtime per task)
Re-running a previous workflow with modified parameters
Sharing run details with collaborators outside your APGAP project

When you click into a pipeline run from APGAP, you’re typically taken straight into the Seqera UI for that run. You won’t usually need to log into Seqera separately — the integration handles authentication transparently for users with the appropriate APGAP roles.

Provisioning behavior#

When you create a new APGAP project, provisioning the Seqera workspace takes a few minutes — typically 5 to 15. During this window, the project exists in APGAP and you can navigate to it, but the Pipelines section will be empty.

What you should see:

Project created → Pipelines section empty for a few minutes
Provisioning completes → Pipelines section populates with the configured pipelines

What to do if it doesn’t:

If the Pipelines section is still empty after 15 minutes, the Seqera workspace provisioning likely failed. The most common causes are GCP quota issues, transient API failures, or IAM permission misconfigurations — all of which are recoverable by a Platform Admin.

If you hit this:

Note the project name and approximate time of creation
Contact your Platform Admin and let them know the Pipelines section is not appearing
The Platform Admin can check the Cloud Build logs in the GCP console, fix the underlying issue, and re-trigger provisioning

While provisioning is being recovered, the rest of APGAP continues to work normally — you can still upload files to the lab, add metadata, and assemble Analytical Datasets. Only the pipeline launching for this specific project is affected. Once the workspace is provisioned, your existing dataset is ready to go.

If Seqera Platform itself is experiencing an outage (rare, but possible), file uploads and metadata management still work normally — only pipeline launching is affected. APGAP and Seqera are loosely coupled, so problems on one side don’t take down the other.