All tutorials
WorkspacePartial

G-protein Design-around

/g-protein-design-around

TL;DR

Technical triage for alternative pseudotyping glycoproteins against VSV-G / Cocal-G risk anchors.

Use it / Skip it

The page ranks candidate envelopes by sequence similarity, equivalent-position mapping, and technical risk so wet-lab and IP review can focus on a smaller set.

Use when

A lentiviral envelope program needs alternatives or a 354-position risk discussion must be made reproducible.

Don't use for

Do not use it as legal advice, freedom-to-operate conclusion, or final packaging protocol.

Inputs

Candidate FASTA
One or more candidate glycoprotein sequences.
References
VSV-G and Cocal-G references; defaults can be fetched where configured.
Optional UniProt query
Exploratory recall for candidate sequence collection.

Outputs

assessments.csv
Ranked spreadsheet with identity, equivalent residue, risk band, score, and reasons.
assessments.json
Machine-readable full result.
manifest.json/report
Input summary and technical-triage caveats.

Walkthrough

  1. 1. Start with a curated FASTA when possible

    Start with a curated FASTA when possible. Broad search results often need manual cleanup.

  2. 2. Run the assessment from the web page or CLI and wait for persisted artifacts

    Run the assessment from the web page or CLI and wait for persisted artifacts.

  3. 3. Read VSV-G identity and 354-equivalent mapping together

    Read VSV-G identity and 354-equivalent mapping together.

  4. 4. Shortlist candidates for structural review, packaging assay, and IP counsel review

    Shortlist candidates for structural review, packaging assay, and IP counsel review.

Under the hood

  • Core implementation lives under pipeline/g_protein_design_around.
  • Biopython PairwiseAligner supports identity and reference-position mapping.
  • The web surface persists jobs and exposes CSV/JSON/report artifacts.

Worked example

Alternative envelope screen
Curate 30 candidate glycoproteins, run the assessment, sort by design-around score, then take the top 10-15 into structure prediction and wet-lab packaging planning.

Pitfalls

  • Below 85% identity is not automatically safe.
  • Very low identity may be legally distant but biologically unusable.
  • Alignment-based equivalent-position mapping needs manual review for distant homologs.

See also