Compare CRISPR Modalities#

This guide helps you choose between CRISPR modalities for your application.

Modality Overview#

Modality	Use Case	Reversible?	Complexity
CRISPRa	Gene activation	Yes	Low
CRISPRi	Gene repression	Yes	Low
Knockout	Gene disruption	No	Low
Prime edit	Precise edits	No	High
Base edit	Single nucleotide changes	No	Medium

Quick Decision Guide#

Want to increase gene expression?

Use CRISPRa

Want to decrease gene expression?

Use CRISPRi (reversible) or Knockout (permanent)

Need a specific mutation?

Single base change: Base editing
Larger change: Prime editing

Gene therapy for haploinsufficiency?

CRISPRa (restore expression from remaining allele)

Comparing CRISPRa vs CRISPRi#

from phaselab.crispr import design_guides, design_crispri_guides

# Same target, different approaches
sequence = "..."  # Target gene promoter
tss_index = 200

# CRISPRa: Activate gene
crispra = design_guides(sequence, tss_index)
print(f"CRISPRa candidates: {len(crispra)}")
print(f"Best CRISPRa: {crispra.iloc[0]['combined_score']:.3f}")

# CRISPRi: Repress gene
crispri = design_crispri_guides(sequence, tss_index)
print(f"CRISPRi candidates: {len(crispri)}")
print(f"Best CRISPRi: {crispri.iloc[0]['combined_score']:.3f}")

Knockout vs CRISPRi#

from phaselab.crispr import design_crispri_guides, design_knockout_guides

# CRISPRi: Reversible repression
crispri = design_crispri_guides(sequence, tss_index)

# Knockout: Permanent disruption
knockout = design_knockout_guides(
    sequence,
    exon_start=500,
    exon_end=800
)

print("CRISPRi advantages:")
print("- Reversible (remove dCas9 to restore)")
print("- No DNA damage")
print("- Tunable (VP64/KRAB fusion strength)")

print("Knockout advantages:")
print("- Complete loss-of-function")
print("- Permanent effect")
print("- Well-established protocols")

Base Editing vs Prime Editing#

from phaselab.crispr import design_base_edit_guides, design_prime_edit

target_position = 150
sequence = "..."

# Base editing: Single base change (A->G or C->T)
# Check if target base is editable
target_base = sequence[target_position]

if target_base == 'A':
    # ABE can convert A->G
    abe_guides = design_base_edit_guides(
        sequence,
        target_position=target_position,
        target_base='A',
        editor_type='ABE'
    )
    print(f"ABE candidates: {len(abe_guides)}")

elif target_base == 'C':
    # CBE can convert C->T
    cbe_guides = design_base_edit_guides(
        sequence,
        target_position=target_position,
        target_base='C',
        editor_type='CBE'
    )
    print(f"CBE candidates: {len(cbe_guides)}")

else:
    print("Base editing not applicable for this base")

# Prime editing: Any edit type
prime = design_prime_edit(
    sequence,
    edit_position=target_position,
    edit_type='substitution',
    new_base='G'
)
print(f"Prime editing pegRNA designed")

Decision Matrix#

def recommend_modality(
    goal: str,
    reversible: bool = False,
    specific_mutation: str = None
) -> str:
    """Recommend CRISPR modality based on requirements."""

    if goal == "increase_expression":
        return "CRISPRa"

    elif goal == "decrease_expression":
        return "CRISPRi" if reversible else "Knockout"

    elif goal == "specific_mutation":
        if specific_mutation in ['A>G', 'T>C']:
            return "ABE (base editing)"
        elif specific_mutation in ['C>T', 'G>A']:
            return "CBE (base editing)"
        else:
            return "Prime editing"

    elif goal == "gene_disruption":
        return "Knockout"

    else:
        return "Consult literature for your specific case"

# Examples
print(recommend_modality("increase_expression"))  # CRISPRa
print(recommend_modality("decrease_expression", reversible=True))  # CRISPRi
print(recommend_modality("specific_mutation", specific_mutation="A>G"))  # ABE

Full Comparison Pipeline#

from phaselab.crispr import (
    design_guides,
    design_crispri_guides,
    design_knockout_guides,
    compute_coherence_batch,
)
import pandas as pd

# Design for all modalities
crispra = design_guides(sequence, tss_index).head(5)
crispra['modality'] = 'CRISPRa'

crispri = design_crispri_guides(sequence, tss_index).head(5)
crispri['modality'] = 'CRISPRi'

knockout = design_knockout_guides(sequence, 500, 800).head(5)
knockout['modality'] = 'Knockout'

# Combine and compare
all_guides = pd.concat([crispra, crispri, knockout], ignore_index=True)

# Add quantum coherence
all_guides['quantum_coherence'] = compute_coherence_batch(
    all_guides['sequence'].tolist(),
    mode="quantum"
)

# Compare by modality
comparison = all_guides.groupby('modality').agg({
    'combined_score': 'mean',
    'quantum_coherence': 'mean',
    'gc_content': 'mean'
})
print(comparison)