Skip to content
MyImmune Docs

Quick Start

This guide walks you through submitting a Clustering pipeline job — from preparing your cohort files to downloading results.

What you’ll need

Section titled “What you’ll need”
  • Access to the MyImmune web app: https://app.svc.myimmune.ai/dashboard
  • A cohort TSV listing your subjects and their AIRR file paths
  • Per-subject AIRR repertoire files, zipped into a single archive

Prepare your input files

Section titled “Prepare your input files”

1. Cohort TSV

Section titled “1. Cohort TSV”

Create a tab-separated file with one row per subject:

subject_id  airr_file_path
S001  S001_airr.zip
S002  S002_airr.zip
S003  S003_airr.zip

Save it as cohort.tsv.

2. Prepare the zip file

Section titled “2. Prepare the zip file”

Zip all per-subject AIRR TSV files along with the cohort file into a single archive:

Terminal window
zip airr_files.zip S001_airr.tsv S002_airr.tsv S003_airr.tsv cohort.tsv

Each AIRR TSV must have at minimum: sequence_id, junction_aa, v_call, j_call. See Data Formats for the full spec.

Submit the job

Section titled “Submit the job”

  1. Log in

    Open the MyImmune web app and sign in.

  2. Open the Clustering Pipeline

    From the Dashboard, click Clustering Pipeline → Create New Job.

  3. Fill in the form

    FieldValue
    Job namee.g., my-first-clustering-run
    Cohort TSVname of the cohort file: cohort.tsv
    Sequence identity (sid)0.85 (default — good starting point)
    Coverage (cov)0.85 (default)
    Generate HTML report✓ enabled
    Advanced Configkeep the default values
    AIRR files zipUpload airr_files.zip

    Leave everything else at defaults for your first run.

  4. Submit

    Click Submit. The job appears in the job list with status queued.

  5. Monitor progress

    If the job is running you can see the status as running. If it has completed successfully you will see the status as success.

  6. Download results

    Once status shows success, the action menu lets you download the results.

What you get

Section titled “What you get”
FileWhat it tells you
cluster_list.tsvEvery cluster — which sequences and subjects belong to it
subject_cluster_map.tsvPer-subject cluster membership for downstream analysis
hierarchical_tree.nwkNewick tree — load in iTOL or FigTree for visualization
cluster_report.htmlInteractive dendrogram + heatmap — open in any browser

What happened behind the scenes

Section titled “What happened behind the scenes”

Each subject’s sequences were compared against themselves at the configured sid/cov thresholds to build subgraphs. Those subgraphs were then clustered hierarchically using the a metric.

See the full Clustering Pipeline reference for all parameters and troubleshooting tips.

Next steps

Section titled “Next steps”
  • Stratification Pipeline — add a response column to your cohort and run LOOCV classification on the clusters you just generated
  • Adjust thresholds — lower sid/cov for broader clusters, or enable auto-tune