| name | experiment-protocol-designer |
| description | Designs rigorous biological experiment protocols with power analysis, control strategy, randomization plan, and step-by-step methodology. Outputs a structured protocol document with exact reagent concentrations, timing, and sample size justification. |
| allowed-tools | Read, Write, WebSearch |
| effort | high |
Experiment Protocol Designer
When to activate
When planning any new biological experiment — before touching a pipette. Use for in-vitro assays, animal studies, cell culture experiments, molecular biology protocols, and field studies.
When NOT to use
Skip for routine lab maintenance tasks, reagent preparation without experimental context, or literature-only reviews without experimental components.
Instructions
- Define the research question: State the hypothesis (H₀ and H₁), primary outcome measure, and expected effect size.
- Power analysis: Calculate minimum sample size using:
- Significance level (α = 0.05 default)
- Power (1-β = 0.80 default)
- Expected effect size (Cohen's d for continuous, OR for categorical)
- Specify the statistical test planned
- Control strategy:
- Positive control: Known response condition
- Negative control: Vehicle/mock treatment
- Technical replicates: Minimum 3 per biological replicate
- Biological replicates: Defined by power analysis
- Randomization: Method (block, stratified, simple), unit of randomization, blinding approach
- Protocol steps: Numbered, with exact:
- Reagents (supplier, catalog number, concentration)
- Equipment (model, settings)
- Timing (incubation duration, temperature)
- Volume/mass measurements with units
- Data collection plan: What to measure, when, instrument, format
- Statistical analysis plan: Test selection, software, correction method
Output Format
PROTOCOL: [Title]
HYPOTHESIS: H₀: [...] H₁: [...]
PRIMARY OUTCOME: [measure + units]
SAMPLE SIZE: n=[X] per group (power analysis: α=0.05, β=0.20, effect size d=[X])
CONTROLS:
Positive: [...]
Negative: [...]
Technical replicates: [X]
RANDOMIZATION: [method] | BLINDING: [method]
MATERIALS:
- [Reagent] ([Supplier], Cat#[number], [concentration])
- [Equipment] ([Model], [settings])
PROCEDURE:
1. [Step with exact conditions]
2. [...]
DATA COLLECTION:
| Timepoint | Measurement | Instrument | Format |
|-----------|-------------|------------|--------|
ANALYSIS PLAN: [Test] with [correction], software: [name]
Example
Request: "Design an experiment to test if compound X inhibits tumor growth in mice"
PROTOCOL: Compound X Anti-Tumor Efficacy Study
HYPOTHESIS: H₀: Compound X has no effect on tumor volume H₁: Compound X reduces tumor volume by ≥40%
PRIMARY OUTCOME: Tumor volume (mm³) measured by caliper at day 21
SAMPLE SIZE: n=12 per group (power analysis: α=0.05, β=0.20, effect size d=1.2 based on pilot)
CONTROLS:
Positive: Doxorubicin 5mg/kg IP q3d
Negative: Vehicle (0.5% CMC in PBS)
Technical replicates: 3 caliper measurements per timepoint
RANDOMIZATION: Block randomization by tumor size at enrollment | BLINDING: Caliper operator blinded to group
MATERIALS:
- Compound X (synthesized in-house, ≥95% purity, 50mg/kg PO)
- Doxorubicin (Sigma, Cat#D1515, 5mg/kg IP)
- Digital caliper (Mitutoyo 500-196-30, ±0.02mm)
PROCEDURE:
1. Implant 1×10⁶ cells subcutaneously in right flank (Day 0)
2. When tumors reach 100-150mm³, randomize to groups (Day 7)
3. Dose daily × 14 days (Day 7-20)
4. Measure tumor volume every 3 days (V = L×W²×0.5)
DATA COLLECTION:
| Timepoint | Measurement | Instrument | Format |
|-----------|-------------|------------|--------|
| D7,10,13,16,20 | Tumor volume | Caliper | mm³ |
| D20 | Tumor weight | Balance | grams |
ANALYSIS PLAN: One-way ANOVA with Dunnett's post-hoc, GraphPad Prism 10