bio-crispr-screens-combinatorial-screens

Imported from PKU-YuanGroup/OpenAI4S (skills/bioskills/bio-crispr-screens-combinatorial-screens/SKILL.md). Install upstream with npx skills add PKU-YuanGroup/OpenAI4S --skill bio-crispr-screens-combinatorial-screens. Copyright stays with the author.

Version Compatibility

Reference examples tested with: MAGeCK 0.5.9+ (for MLE with interaction terms), Inzolia library annotation (Esmaeili Anvar 2024), pandas 2.2+, numpy 1.26+, scipy 1.12+, matplotlib 3.8+.

Before using code patterns, verify installed versions match. If versions differ:

• CLI: mageck --version; mageck mle --help
• For Cas12a libraries: verify against published Inzolia / in4mer / Big Papi annotations

If code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API rather than retrying.

Combinatorial CRISPR Screen Analysis

"Run a combinatorial CRISPR screen to find synthetic-lethal interactions" -> Design a paired or multiplex library, screen for double-knockout fitness, score per-pair genetic interaction (GI = observed_double - expected_additive), and identify synthetic-lethal (negative GI) and synthetic-rescue (positive GI) interactions.

• CLI: mageck mle with explicit interaction terms for paired-Cas9 (Big Papi-style)
• Python: custom GI scoring for Cas12a multiplex (in4mer / Inzolia)
• Modality: enCas12a / LbCas12a single-array multiplex (preferred for paralog screens)

Combinatorial Architecture Decision Tree

Fails when:

• Dual-sgRNA constructs built from repeated U6/tracr elements: lentiviral recombination collapses them to a single perturbation
• Cas12a screens analyzed as Cas9 screens: MAGeCK normalization fails because Cas12a has different cut profile
• in4mer 4-guide arrays without all-singleton controls: GI scoring requires single-gene baselines

Cas9 vs Cas12a for Multiplex

Recommendation: For modern paralog screens, use Cas12a multiplex with the Inzolia library. It is ~30% smaller than a typical monogenic Cas9 library while additionally covering ~4,000 paralog pairs (Esmaeili Anvar 2024), which makes it more cost-effective at genome scale.

The Paralog Buffering Phenomenon

Dede et al 2020 Genome Biol 21:262 showed that a large share of constitutively expressed genes are never scored as essential in any Cas9 single-KO fitness screen (3,032 of 7,282; 42%), and that these never-essentials are strongly enriched for paralogs. The reason: gene paralogs perform redundant essential functions. Loss of one paralog is buffered by the other; only loss of both creates the essentiality phenotype.

Quantified impact: 24 synthetic-lethal paralog pairs identified in Dede 2020 across 3 cell lines; 19 of 24 (79%) reproduce in >=2 lines, 14 of 24 (58%) in all 3. These pairs were not findable by single-gene Cas9 screens, requiring combinatorial methodology.

Examples:

• MAPK1 (ERK2) + MAPK3 (ERK1): ERK family redundancy in proliferation
• PIK3CA + PIK3CB: PI3K alpha/beta redundancy
• AKT1 + AKT2: AKT family redundancy
• HSP90AA1 + HSP90AB1: HSP90 alpha/beta redundancy
• STAG1 + STAG2: Cohesion complex paralogs

Each is buffered: loss of one is tolerated; loss of both is lethal.

Genetic Interaction (GI) Scoring

Goal: Identify pairs where the double-knockout fitness differs from the additive expectation.

Approach: From per-pair and per-singleton fitness data, compute GI = observed_double_LFC - (single_A_LFC + single_B_LFC). Synthetic lethal: GI < threshold (more depleted than additive). Synthetic rescue: GI > threshold (less depleted than additive).

import pandas as pd
import numpy as np
from scipy.stats import zscore

def gi_score(paired_lfc_df, single_lfc_df):
    '''Score genetic interactions from paired vs single LFCs.

    paired_lfc_df: rows = paired-KO; columns = ['gene_A', 'gene_B', 'paired_lfc']
    single_lfc_df: rows = single-KO; columns = ['gene', 'single_lfc']
    '''
    single = dict(zip(single_lfc_df['gene'], single_lfc_df['single_lfc']))
    df = paired_lfc_df.copy()
    df['single_A_lfc'] = df['gene_A'].map(single)
    df['single_B_lfc'] = df['gene_B'].map(single)
    df['expected_additive'] = df['single_A_lfc'] + df['single_B_lfc']
    df['gi_score'] = df['paired_lfc'] - df['expected_additive']
    df = df.dropna(subset=['gi_score'])          # a single missing singleton would NaN every z-score
    df['gi_z'] = zscore(df['gi_score'])
    df['gi_class'] = np.where(df['gi_z'] < -2, 'synthetic_lethal',
                                np.where(df['gi_z'] > 2, 'synthetic_rescue', 'no_interaction'))
    return df.sort_values('gi_z')

Interpretation:

• GI z-score < -2: Synthetic lethal (double-KO more lethal than expected) -- candidate drug target combinations
• GI z-score > 2: Synthetic rescue (double-KO less lethal than expected) -- compensatory pathway / paradoxical hit
• GI z-score -1 to 1: No interaction; effects are additive

Run Combinatorial Screen Analysis (MAGeCK MLE with Interaction Indicator)

Goal: Use MAGeCK MLE to estimate the effect of each gene independently and the additional effect when both genes are simultaneously perturbed.

Approach: Design matrix encodes single-A, single-B, double-AB conditions; the interaction column is set to 1 only for double-KO samples. The resulting beta for that column captures the extra effect beyond the sum of single-gene betas. Note: MAGeCK MLE does not natively perform a formal interaction-significance test, but the interaction|beta and |fdr columns serve as the GI estimate; for formal interaction testing, compute GI = observed_double_lfc - (single_A_lfc + single_B_lfc) explicitly (see GI scoring section below).

# Design matrix encoding double-KO as a separate "interaction" indicator
# Conditions: NT (control), A_KO, B_KO, A_B_KO
cat > combo_design.txt <<EOF
Samples         baseline    geneA       geneB       interaction
NT_r1           1           0           0           0
NT_r2           1           0           0           0
A_r1            1           1           0           0
A_r2            1           1           0           0
B_r1            1           0           1           0
B_r2            1           0           1           0
AB_r1           1           1           1           1
AB_r2           1           1           1           1
EOF

mageck mle \
    --count-table combo_counts.txt \
    --design-matrix combo_design.txt \
    --output-prefix combo_mle

# Output: per-gene beta scores per design column
# The "interaction" column beta captures additional joint effect beyond additive

Interpretation of MAGeCK MLE output:

A significantly negative interaction|beta is synthetic lethal; positive is synthetic rescue. For formal GI hypothesis testing, prefer the explicit GI scoring approach (next section) over MAGeCK MLE interpretation, since MAGeCK MLE does not validate the additive null.

Inzolia / in4mer 4-Guide Array Analysis

Esmaeili Anvar 2024 Nat Commun 15:3577 introduced in4mer, a Cas12a multiplex architecture where each array contains 4 guides processed by Cas12a's intrinsic crRNA-processing activity. The Inzolia library is the canonical implementation, covering the protein-coding genome plus ~4,435 paralog pairs.

Library design:

• 4 guides per cassette (Cas12a single-transcript array)
• Per pair: 2 arrays carrying 2 guides per gene, with the guides presented in different order across the two arrays
• Includes singleton controls: each single gene is covered by 2 four-guide arrays (4 guides per gene, order swapped)
• ~49,000 total arrays covering 19,687 genes, ~4,435 paralog pairs, 376 triples, and 100 quads

# Per-pair analysis from in4mer screen
def in4mer_pair_analysis(paired_counts_df, gene_pairs, value_cols):
    '''Aggregate cassette-level counts to per-pair statistics.
    paired_counts_df: rows = cassettes, with a cassette_id COLUMN (reset_index first if it is the index).
    gene_pairs: DataFrame with cassette_id and gene_A, gene_B columns.
    value_cols: the numeric sample/LFC columns to aggregate.
    '''
    merged = paired_counts_df.merge(gene_pairs, on='cassette_id')
    return merged.groupby(['gene_A', 'gene_B'])[value_cols].agg(['mean', 'std', 'count'])

Failure Modes

Dual-sgRNA construct recombines in the lentiviral vector

Trigger: A dual-sgRNA construct built from repeated elements -- two copies of the U6 promoter, or two copies of the SpCas9 tracrRNA scaffold. Mechanism: Najm 2018 reports that repetitive elements in lentiviral vectors, including the U6 promoter and multiple copies of the tracrRNA sequence, drive high levels of recombination and reduce combinatorial screen efficiency. Big Papi avoids this by pairing two orthologous enzymes (SaCas9 + SpCas9), whose scaffolds differ, and expressing the two sgRNAs from distinct U6 and H1 promoters. Symptom: Constructs collapse to a single perturbation; measured GI scores are diluted toward zero. Fix: Use the pPapi architecture (orthologous Cas9s, U6 + H1) rather than duplicated U6/tracr elements; verify construct integrity by amplicon sequencing of clones.

Cas12a screen with low editing efficiency

Trigger: Cas12a less efficient than Cas9 at some loci; some guides in the 4-guide array don't cut. Mechanism: Cas12a editing rate varies by sequence context; some loci edit at <30%. Symptom: Specific pairs missing expected effects despite cassette presence. Fix: Pilot Cas12a efficiency at the loci before full screen; use enCas12a (enhanced) variant; for known low-efficiency loci, supplement with Cas9.

GI scoring without singletons

Trigger: Library lacks single-gene controls (only paired knockouts). Mechanism: GI = paired - expected_additive requires single-gene LFC; without them, expected cannot be computed. Symptom: Cannot score GI; only paired LFCs available. Fix: Design library to include singletons (place gene A with 3 placeholder guides; gene B with 3 placeholders); re-run with full design.

Single-gene LFCs from different cell line

Trigger: Using public single-gene LFCs (e.g., DepMap) as the baseline for paired-screen GI scoring. Mechanism: Single-gene effects are cell-line specific; using HCT116 single-gene LFCs to score K562 paired-screen GIs is invalid. Symptom: GI scores look noisy; many false positives. Fix: Include singleton controls in the screen; or use cell-line-matched DepMap data.

Confounding cell-cycle / proliferation in GI scoring

Trigger: Paired KO of two cell-cycle-impacting genes; the double-effect saturates cell cycle. Mechanism: If A_KO causes 50% growth arrest and B_KO causes 50%, the combined 75% arrest is already saturating proliferation; additive expectation overestimates double-effect, generating false "synthetic-rescue." Symptom: GI scores positive for pairs of essential cell-cycle genes; biologically unexpected. Fix: Use log-space (LFC) GI scoring rather than linear; saturation is less severe in log-space. Alternative: model with logistic / saturable response curve.

Library skew amplifying noise

Trigger: Inzolia library has uneven cassette representation; some pairs at 10x lower coverage than others. Mechanism: Standard library QC (Gini, skew) applies; low-coverage cassettes yield noisier LFCs. Symptom: GI z-scores vary 2-3x across cassettes targeting the same pair. Fix: Standard library QC; for low-coverage pairs, aggregate fewer cassettes but with more sequencing depth; or drop low-coverage pairs from analysis.

Cross-Modality Validation

For high-stakes synthetic-lethal hits (drug-target nomination), validate by:

1. Orthogonal chemistry: Re-validate with Cas9 if Cas12a, or vice versa
2. Arrayed validation: Single-knock-out arrayed setup with same cell line; quantify proliferation
3. CRISPRi orthogonal: Use dCas9-KRAB to confirm knockdown phenotype (no DNA damage)
4. Pharmacological: Inhibit paralog with drug; confirms target accessibility for drug development

Quantitative Thresholds

Common Errors

References

• Najm FJ et al. 2018. Nat Biotechnol 36:179. Big Papi paired-Cas9 platform.
• DeWeirdt PC et al. 2021. Nat Biotechnol 39:94. enAsCas12a multiplex.
• Esmaeili Anvar N et al. 2024. Nat Commun 15:3577. in4mer / Inzolia paralog library.
• Dede M et al. 2020. Genome Biol 21:262. Paralog buffering in Cas9 screens.
• Thompson NA et al. 2021. Nat Commun 12:1302. Combinatorial CRISPR screen identifying paralog fitness effects.
• Boettcher M et al. 2018. Nat Biotechnol 36:170. Dual CRISPR activation + knockout directional genetic-interaction screen.
• Horlbeck MA et al. 2018. Cell 174:953-967. CRISPRi combinatorial genetic-interaction map; paralog buffering as a co-essentiality pattern was characterized more directly in Dede 2020 Genome Biol 21:262 and Gonatopoulos-Pournatzis 2020 Nat Biotechnol 38:638.

Related Skills

• crispr-screens/library-design - Inzolia / in4mer / Big Papi library design
• crispr-screens/screen-qc - Library QC including cassette skew
• crispr-screens/mageck-analysis - MAGeCK MLE with interaction terms
• crispr-screens/hit-calling - Cross-method analysis of combinatorial data
• crispr-screens/perturb-seq-analysis - Combinatorial Perturb-seq
• crispr-screens/copy-number-correction - Pre-correction for cancer-line combinatorial screens
• crispr-screens/in-vivo-screens - In-vivo paralog screens
• pathway-analysis/go-enrichment - Functional analysis of GI clusters