Acceptance Criteria and Submission

You have a hypothesis and a protocol. Before submitting, you need to define what success looks like and set your constraints.

Why Acceptance Criteria Matter

Acceptance criteria are your pre-registered definition of success. They:

Prevent moving goalposts after seeing data
Make disputes objective (did results meet criteria or not?)
Force you to think through what you actually need
Enable automated validation by AI pipelines

Define these before the experiment runs.

Defining Success Conditions

Each success condition has three parts:

Metric: What you're measuring
Operator: How to compare (greater than, less than, between, etc.)
Threshold: The value that constitutes success

Example: Enzyme Inhibition

{
  "acceptance_criteria": {
    "success_conditions": [
      {
        "metric": "IC50",
        "operator": "lte",
        "threshold": 10,
        "unit": "μM",
        "description": "Compound must inhibit enzyme with IC50 ≤ 10μM"
      },
      {
        "metric": "Maximum inhibition",
        "operator": "gte",
        "threshold": 80,
        "unit": "%",
        "description": "Must achieve at least 80% inhibition at highest concentration"
      }
    ]
  }
}

Available Operators

Operator	Meaning	Use Case
`gt`	Greater than	"Effect must be >X"
`gte`	Greater than or equal	"Effect must be ≥X"
`lt`	Less than	"IC50 must be <X"
`lte`	Less than or equal	"MIC must be ≤X"
`eq`	Equal to	Rarely used (too strict)
`neq`	Not equal to	"Must differ from control"
`between`	Within range	"pH must be 6.5-7.5"
`not_between`	Outside range	"Must not be in toxic range"

Using Ranges

For between and not_between, use an object threshold:

{
  "metric": "Cell viability",
  "operator": "between",
  "threshold": { "min": 80, "max": 120 },
  "unit": "% of control",
  "description": "Compound must not be cytotoxic (viability 80-120% of control)"
}

Defining Failure Conditions

Failure conditions describe results that would definitively reject your hypothesis:

{
  "failure_conditions": [
    {
      "condition": "No inhibition observed at any concentration up to 1mM",
      "interpretation": "Compound is inactive against this target"
    },
    {
      "condition": "IC50 > 100μM",
      "interpretation": "Compound has weak activity, not suitable for development"
    }
  ]
}

Statistical Requirements

For quantitative experiments, specify your statistical standards:

{
  "statistical_requirements": {
    "significance_level": 0.05,
    "statistical_test": "t_test",
    "power": 0.8
  }
}

Test	Use When
`t_test`	Comparing two groups, continuous data, normal distribution
`anova`	Comparing 3+ groups
`mann_whitney`	Comparing two groups, non-normal data
`chi_square`	Categorical outcomes
`fisher_exact`	Categorical outcomes, small sample sizes

Setting Constraints

Budget

{
  "constraints": {
    "budget_max_usd": 500,
    "budget_flexibility": "flexible_25"
  }
}

Flexibility options:

strict: Hard cap, no exceptions
flexible_10: Can go up to 10% over if needed
flexible_25: Can go up to 25% over
flexible_50: Can go up to 50% over

Turnaround

{
  "turnaround_days": 14,
  "turnaround_priority": "standard"
}

Priority levels:

standard: Normal queue, best pricing
expedited: Faster matching, ~20% premium
urgent: Top priority, ~50% premium

Privacy

{
  "privacy": "open"
}

Option	Description	Premium
`open`	Results public immediately	None
`delayed_6mo`	Private for 6 months, then public	~10%
`delayed_12mo`	Private for 12 months, then public	~15%
`private`	Permanently private	~25-50%

Open science is the default. You retain IP on your hypothesis regardless of privacy setting.

Submitting via API

Validate First (Recommended)

curl -X POST https://api.litmus.science/validate \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d @experiment.json

Response:

{
  "valid": true,
  "errors": [],
  "warnings": [
    {
      "path": "protocol.replicates.biological_replicates",
      "code": "low_replicates",
      "message": "2 biological replicates may not provide sufficient statistical power"
    }
  ],
  "safety_flags": []
}

Get Cost Estimate

curl -X POST https://api.litmus.science/estimate \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d @experiment.json

Submit

curl -X POST https://api.litmus.science/experiments \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d @experiment.json

Response:

{
  "experiment_id": "exp_7a8b9c0d-1234-5678-9abc-def012345678",
  "status": "open",
  "created_at": "2026-01-15T10:30:00Z",
  "estimated_cost_usd": 250,
  "estimated_turnaround_days": 14
}

What Happens Next

Open: Your experiment is visible to qualified operators
Claimed: An operator accepts the job (you're notified)
In Progress: Operator executes the protocol
Results Submitted: You review results against acceptance criteria
Approved/Disputed: You approve (payment released) or dispute

After Results: Approval or Dispute

Approving Results

If results meet your acceptance criteria:

curl -X POST https://api.litmus.science/experiments/exp_xxx/approve \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"rating": 5, "feedback": "Excellent documentation, clear results."}'

Disputing Results

If results don't meet criteria or there are quality issues:

curl -X POST https://api.litmus.science/experiments/exp_xxx/dispute \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "reason": "results_incomplete",
    "description": "Protocol specified 3 biological replicates but only 2 were performed."
  }'

Disputes are resolved by Litmus based on whether delivered work meets the stated acceptance criteria.

Tips for Successful Experiments

Start small: First experiment? Keep it simple. One variable, standard template, low budget.
Be specific in acceptance criteria: Vague criteria lead to disputes. "Must show activity" is bad. "IC50 ≤ 10μM" is good.
Budget realistically: Use the estimate endpoint. Lowballing your budget reduces operator interest.
Provide quality materials: If you're shipping compounds, ensure purity and proper storage.
Communicate: If an operator has questions, respond promptly.
Review results carefully: Check that the protocol was followed, not just the final numbers.

Quick Reference: Submission Checklist

Protocol Selection Overview