Quick start

This page is a fast path for first-time users. Install RAFT first using Installing RAFT, then run the built-in LENS demonstration to confirm your environment is working end to end.

Before you start

LENS is a large workflow. It is worth confirming these before your first run rather than discovering them partway through a download.

What

What to expect

Disk

The human reference set is roughly 130 GB, and container images add about 27 GB across ~50 images. Both are one-time downloads shared across projects. Each project then needs its own working space for intermediates on top of that.

Memory

Plan for a node that can give a single task at least 40 GB. RNA alignment is the peak: STAR loads the genome index into RAM, so its footprint is set by the reference rather than by how much data you have.

Time

Roughly 7-12 hours for one patient with WES tumor/normal plus tumor RNA-seq. This is a working estimate rather than a benchmark, and it shifts with sequencing depth, node size, and how much of the workflow you enable.

Software

Nextflow 26.04.6, RAFT v1.5.0 or newer, Git, and Apptainer or Singularity. See Installing RAFT. Installing these usually needs administrator access; on a shared cluster, check module avail apptainer or ask your system administrator.

Optional

A NetMHCpan academic licence, only if you want NetMHC-derived columns (:doc: netmhc_setup). LENS uses MHCflurry by default and does not require it.

Install RAFT before continuing. Installing RAFT covers RAFT, its system dependencies, and the workspace the demonstration runs in.

Running the demonstration

The demonstration runs end to end on one melanoma patient (Pt01) from Genomic and Transcriptomic Features of Response to Anti-PD-1 Therapy in Metastatic Melanoma (Hugo et al., 2016).

cd to your raft/ directory and run:

$ raft run-demo --workflow lens --version 1.9.3

RAFT creates a project called demo-lens-v1.9.3 under raft/projects/, downloads the workflow modules, demo references, FASTQs, and manifest, then runs the workflow.

Once it finishes, see Accessing and reviewing outputs for where the results land and how to review them.

What’s next?

After the demonstration completes: