SMS Therapeutic Pipeline#

This tutorial covers the complete Smith-Magenis Syndrome (SMS) gene therapy assessment pipeline in PhaseLab.

Overview#

Smith-Magenis Syndrome is caused by RAI1 haploinsufficiency (~50% expression). The SMS trials module provides:

CRISPRa RAI1 activation - Boost remaining allele
CRISPRi modifier suppression - Target PER1, CRY1, CLOCK
Circadian rescue simulation - Predict sleep/wake improvement
Delivery assessment - AAV feasibility for CNS

Quick Start#

from phaselab.trials.sms import SMSPipeline, SMSTrialConfig

# Configure pipeline
config = SMSTrialConfig(
    therapeutic_window=(0.70, 1.10),  # 70-110% of normal RAI1
    optimal_expression=0.80,           # Target 80%
    verbose=True,
)

# Run full assessment
pipeline = SMSPipeline(config=config)
result = pipeline.run_full_pipeline()

# Get GO/NO-GO decision
print(f"Overall: {result.overall_go_nogo}")
print(f"Claim level: {result.overall_claim_level}")

Understanding Results#

# Individual trial results
print(f"CRISPRa candidates: {result.crispra_result.n_candidates}")
print(f"Circadian rescue: {result.circadian_result.metrics['rescue_status']}")
print(f"Delivery feasible: {result.delivery_result.metrics['delivery_feasibility']}")

# Get falsification tests for wet lab
for test in result.falsification_tests:
    print(f"Test {test['id']}: {test['name']}")
    print(f"  Failure condition: {test['failure_condition']}")

Individual Trials#

CRISPRa RAI1 Activation#

from phaselab.trials.sms import run_sms_crispra_trial

result = run_sms_crispra_trial(
    promoter_sequence=rai1_promoter,  # Optional: uses default if not provided
    config=config,
)

print(f"Status: {result.status}")
print(f"Candidates: {result.n_candidates}")

if result.best_candidate:
    print(f"Best guide: {result.best_candidate['sequence']}")
    print(f"Position: TSS{result.best_candidate['position']:+d}")

CRISPRi Modifier Suppression#

from phaselab.trials.sms import run_sms_crispri_trial

# Target circadian modifier genes
for target in ["PER1", "CRY1", "CLOCK"]:
    result = run_sms_crispri_trial(
        target_gene=target,
        config=config,
    )

    print(f"{target}: {result.n_candidates} candidates")
    if result.best_candidate:
        supp = result.best_candidate.get('expected_suppression', 'N/A')
        print(f"  Expected suppression: {supp}")

Circadian Rescue Simulation#

from phaselab.trials.sms import run_circadian_rescue_simulation

result = run_circadian_rescue_simulation(
    predicted_rai1_expression=0.80,  # After CRISPRa
    config=config,
)

print(f"Rescue status: {result.metrics['rescue_status']}")
print(f"Final R̄: {result.metrics['final_R_bar']:.3f}")
print(f"Sleep quality: {result.metrics['sleep_quality_prediction']}")

AAV Delivery Assessment#

from phaselab.trials.sms import run_delivery_assessment

result = run_delivery_assessment(
    modality="CRISPRa_VP64",
    target_tissue="brain",
    config=config,
)

print(f"Feasibility: {result.metrics['delivery_feasibility']}")
print(f"Payload size: {result.metrics['payload_size']}bp")
print(f"Serotype: {result.best_candidate['recommended_serotype']}")

Falsification Tests#

PhaseLab generates specific tests that would prove predictions wrong:

Falsification Tests#
Test	Name	Failure Condition
A	Ranking Validity	PhaseLab guides don’t outperform random CRISPOR controls
B	Risk Prediction	CAUTION guides don’t fail more than SAFE guides
C	Dosage Prediction	Expression correlation < 0.6
D	UNKNOWN Calibration	UNKNOWN guides don’t fail at random rate

These tests are designed to be run in the wet lab to validate or falsify PhaseLab’s predictions.

Claim Levels#

The pipeline reports uncertainty honestly:

UNKNOWN: Cannot assess reliability
EXPLORATORY: Structural priors only (no tiling data)
CONTEXT_DEPENDENT: Single dataset validation
STRONG_COMPUTATIONAL: Cross-validated predictions

Important

All SMS therapeutic claims require wet lab validation before clinical use. PhaseLab provides computational guidance, not clinical recommendations.

Configuration Options#

config = SMSTrialConfig(
    # Therapeutic parameters
    therapeutic_window=(0.70, 1.10),
    optimal_expression=0.80,
    baseline_expression=0.50,  # SMS patients

    # Pipeline settings
    use_virtual_assay=True,     # Enhanced scoring
    coherence_mode="heuristic", # Fast mode (or "quantum")

    # Simulation parameters
    simulation_hours=120.0,     # 5 days
    n_circadian_trials=5,       # Replicates

    # Output
    verbose=True,
)

Full Example#

from phaselab.trials.sms import (
    SMSPipeline,
    SMSTrialConfig,
)

# Configure
config = SMSTrialConfig(
    therapeutic_window=(0.70, 1.10),
    verbose=True,
)

# Run pipeline
pipeline = SMSPipeline(config=config)
result = pipeline.run_full_pipeline(
    include_modifiers=True,   # CRISPRi trials
    include_editing=True,     # Base/prime editing
)

# Summary
print("=" * 60)
print("SMS GENE THERAPY ASSESSMENT")
print("=" * 60)
print(f"Overall GO/NO-GO: {result.overall_go_nogo}")
print(f"Claim level: {result.overall_claim_level}")
print()

# CRISPRa results
print(f"CRISPRa candidates: {result.crispra_result.n_candidates}")
if result.crispra_result.best_candidate:
    guide = result.crispra_result.best_candidate
    print(f"  Lead guide: {guide['sequence'][:20]}...")
    print(f"  Position: TSS{guide['position']:+d}")

# Circadian prediction
print(f"Circadian rescue: {result.circadian_result.metrics['rescue_status']}")

# Wet lab recommendations
print("\nWet Lab Recommendations:")
for rec in result.wet_lab_recommendations[:5]:
    print(f"  - {rec}")

# Falsification tests
print("\nFalsification Tests:")
for test in result.falsification_tests:
    print(f"  {test['id']}: {test['name']}")

Next Steps#

/reference/trials - Full API reference
Claim Levels - Understanding claim levels
/user_guide/howtos/export_for_wetlab - Export guides for ordering