thebench.bio #1 Ecosystem for the Bio-Tech Industry
All Molecular Biology questions

Question 9 of 9 Advanced Practical

Design a rescue experiment to determine whether a phenotype is caused by the intended genetic perturbation.

Short answer

A convincing rescue restores the expected phenotype by reintroducing the relevant gene product, with controls that rule out overexpression artifacts and unrelated pathway bypass.

Why

Rescue strengthens causality by changing the manipulated factor again: perturbation should create the phenotype and specific restoration should reverse it. However, a rescue construct can create a second, artificial mechanism, especially if its expression or localization differs from the endogenous protein.

How to work it through

  1. Use at least two independent perturbations where feasible.
  2. Introduce a perturbation-resistant wild-type construct at a near-physiological expression level, matched empty vector and a nonfunctional or mechanism-specific mutant.
  3. Quantify expression, localization and the original endpoint under blinded analysis.
  4. Include an orthogonal assay and repeat across biological replicates.
  5. For CRISPR guides, silent changes to the guide-binding region can prevent re-cutting without altering protein sequence.

An example

Guide A and guide B each reduce mitochondrial respiration. guide-resistant wild-type gene restores respiration but a catalytically inactive mutant does not. This supports a requirement for catalytic activity, assuming controls exclude altered growth or cell number.

Where the answer stops holding

Rescue may fail because of expression timing, isoform, stoichiometry or toxicity, not because the original phenotype was off-target. Partial rescue can be informative. Rescue by a paralogue establishes pathway substitutability rather than uniqueness.

A strong answer covers

  • A convincing rescue restores the expected phenotype by reintroducing the relevant gene product, with controls that rule out overexpression artifacts and unrelated pathway bypass
  • Use at least two independent perturbations where feasible

Common mistakes

  • Using massively overexpressed cDNA as decisive proof
  • No empty-vector control
  • Measuring a different phenotype at rescue