Skip to content

Offboarding

When you leave the Roberts Lab — graduating, finishing a postdoc or appointment, wrapping up an undergraduate position, or moving on for any reason — the goal is simple: your work stays usable by the lab, and you leave with copies of everything that is yours to keep.

Use this page as a checklist. Start 2–4 weeks before your last day so nothing depends on access you no longer have. Work through it with Steven, and check items off as you go.

!!! tip "The one-line rule" Everything that matters lives in a GitHub repository under RobertsLab, on the lab servers, or in a shared lab account — not on your laptop, your personal cloud drive, or in your head. If it only exists in one of those places, move it before you go.

1. Data — make sure nothing is stranded

The lab must retain all research data, and you are welcome to keep copies of work you contributed to.

  • [ ] Raw data (sequencing reads, images, instrument output, field/experimental measurements) is on the lab servers and/or in a permanent archive (e.g., NCBI SRA/GEO, OWL/Gannet). Note accession numbers in the relevant repo.
  • [ ] Processed/analyzed data and intermediate files are stored or clearly reproducible from code. Large files belong on lab storage (Gannet/Owl), not in Git — link to them.
  • [ ] Metadata is complete: sample sheets, README files, units, and collection details so someone else can interpret the data without you. See Data Management.
  • [ ] Physical samples (extractions, tissues, slides, plates) are labeled and logged in the Roberts Lab Inventory System, with freezer/box locations recorded. See the Lab Inventory handbook page for related databases and guidance.
  • [ ] Your personal copy: export or download any data you are entitled to keep. Confirm with Steven what you may take and in what form.

2. GitHub — transfer ownership, don't take it with you

All lab project repositories should live in the RobertsLab GitHub organization, not in your personal account.

  • [ ] Move any project repos currently under your personal account into the RobertsLab org (Repo → Settings → Transfer ownership). If a repo can't be transferred, push a complete copy to a RobertsLab repo.
  • [ ] Push your final commits. No uncommitted or unpushed work on your local machine.
  • [ ] Update each repo's README so a new person can pick it up: what the project is, where the data lives, how to run the analysis, and current status.
  • [ ] Open issues / loose ends are written down as GitHub Issues so they aren't lost.
  • [ ] Your lab notebook (GitHub-based) is complete and remains in the org.
  • [ ] After your access is removed, you keep your personal GitHub account — you can keep forks/copies of public work, but the org's copy is the canonical one.

3. Notebooks and code — leave it reproducible

  • [ ] Lab notebook is up to date through your last working day and stays accessible to the lab.
  • [ ] Analysis notebooks (Jupyter .ipynb, R Markdown / Quarto .qmd/.Rmd, scripts) are committed to the relevant repo and run top-to-bottom from the data described in the README.
  • [ ] Environments are captured — environment.yml, renv.lock, requirements.txt, or a container — so the analysis can be re-run later. See Computing Best Practices.
  • [ ] Your personal copies: clone/download the repos and notebooks you want to keep before access is removed.

4. Keys, accounts, and access — hand off, then have it revoked

Make a quick list of everything you can log into for the lab, then work through it.

  • [ ] Server access (Klone/Hyak, Raven, Gannet, Owl, lab workstations): confirm your files are off your personal home directories and onto shared/permanent storage, then have your accounts deactivated.
  • [ ] Shared credentials / API keys / tokens you used (cloud, database, instrument, sequencing-core, service accounts): identify them so they can be rotated. Never email keys in plain text or leave them in committed code — see Data Management.
  • [ ] GitHub: any personal access tokens or deploy keys tied to lab repos are revoked.
  • [ ] Communication & shared accounts: lab Slack, Lab Calendar, shared Google Drive folders, Zotero/citation groups, social or project accounts — transfer anything you alone control, then have your access removed.
  • [ ] Physical access: return keys/fobs, and have building/room and freezer access updated per Lab Safety.
  • [ ] UW accounts: know that your NetID, UW email, and UW-licensed software access typically expire after you leave — move anything you need out of UW-only systems first.

5. Knowledge transfer — so the work continues

  • [ ] Hand off active projects to a named lab member (or to Steven), with a short written summary: status, what's next, where everything lives.
  • [ ] Manuscripts in progress: confirm authorship, share the latest drafts and source files (figures, code, data), and agree on who carries each paper forward. See Writing Scientific Manuscripts and Pubathon.
  • [ ] Document the undocumented: protocols, scripts, or "tricks" only you know go into the repo or Protocols wiki.
  • [ ] Update Project Management boards/issues to reflect reality.

6. Stay connected

  • [ ] Forwarding contact: give Steven a personal (non-UW) email and, if you like, your new affiliation so we can reach you about papers, data questions, and references.
  • [ ] ORCID & affiliations: make sure your ORCID record reflects the lab and your publications.
  • [ ] Letters of recommendation / references: let Steven know if you'll need them — easier to arrange before you go.
  • [ ] Keep in touch. Alumni are part of the lab community; you're welcome to stay in the relevant channels and at lab events.

Final checklist (the short version)

Area Done when…
Data Raw + processed data and metadata are on lab storage/archives; samples are logged; you have your personal copies
GitHub All project repos are under RobertsLab with current READMEs; nothing unpushed
Notebooks/code Notebooks run end-to-end; environments captured; lab notebook complete
Keys & access Shared keys flagged for rotation; server/UW/personal accounts handled; physical keys returned
Knowledge transfer Active projects + manuscripts handed off in writing; authorship confirmed
Stay connected Forwarding email shared; ORCID updated; references arranged

Questions about what you may keep, what must stay, or how to transfer something? Ask Steven before your last day. See also Onboarding, Data Management, and Environment and Expectations.