Genome-Wide CRISPR Guide Design for Liver-Targeted Gene Therapy

Gene Therapy

Genome-Wide CRISPR Guide Design for Liver-Targeted Gene Therapy

Series A gene therapy startup needed high-specificity guide RNAs for an in vivo hepatocyte edit with zero tolerance for off-target cuts at pathogenic loci.

Company

Series A gene therapy startup (8-person team)

Timeline

September 2025

Engagement

CRISPR Guide Design & Off-Target Screening Pipeline

CRISPR Guide Design

10 days: End-to-end delivery
95%+: On-target specificity (top 5)
0: Detected off-target cuts
5/5: First-pass in vitro validation

The Challenge

A Series A gene therapy startup was developing an in vivo CRISPR therapy for a rare liver metabolic disorder. They needed guide RNA sequences optimized for on-target cutting efficiency in hepatocytes while avoiding off-target edits across the human genome. A CRO quoted 8 to 10 weeks and $180K for manual guide design. The startup had no computational biology team and needed sequences before a manufacturing slot reservation deadline.

Business Constraints

Budget: $220K (total computational budget for preclinical phase)
Timeline: Ranked guide sequences in 10 business days
Zero tolerance for off-target cuts at known pathogenic loci

GeneForge Approach

Days 1 to 3: Target Region Analysis and Variant Generation

Input

Target gene locus (GRCh38/hg38 coordinates)
Regulatory constraint data (ClinVar, gnomAD constraint scores)
Prior published guide sequences for the same locus (negative controls)

Methods

500K guide RNA variant generation across target exons and splice sites
On-target activity prediction (DeepCRISPR + proprietary ensemble)
Genome-wide off-target search with up to 3 mismatches (Cas-OFFinder + BLAST)

Output

Top 2,400 guides ranked by on-target activity score
Off-target risk map across all 500K variants

Days 4 to 7: Off-Target Filtering and Tissue Context Optimization

Guides with off-target matches at pathogenic loci or essential genes were eliminated (847 removed). Chromatin accessibility data (ENCODE liver tissue) weighted on-target predictions for hepatocyte-specific activity. RNA secondary structure and manufacturability scoring (synthesis complexity, GC content) filtered to 180 high-confidence guides. PAM diversity analysis ensured backup guides if primary PAM sites showed polymorphism in patient populations.

Days 8 to 10: Ranking and Validation Protocol

Output

Top 20 guide sequences with on-target score, off-target count, and specificity tier
Recommended in vitro validation protocol (T7E1 assay + amplicon sequencing)
Off-target confirmation panel (top 10 predicted off-target sites per guide)
Manufacturing-ready sequence files for top 5 guides

Final Ranked Guide Sequences

Top 5 shown; full list of 20 delivered with manufacturing files.

Top CRISPR guides by on-target activity and off-target safety
Rank	Guide ID	On-Target Score	Off-Target Sites	Specificity	Status
1	GF-g001	0.94	0 critical	0.97	Priority A
2	GF-g002	0.92	0 critical	0.96	Priority A
3	GF-g003	0.91	1 low-risk	0.94	Priority A
4	GF-g004	0.89	0 critical	0.95	Priority A
5	GF-g005	0.88	1 low-risk	0.93	Priority A
6–20	GF-g006–g020	0.82–0.87	0–2 low-risk	0.88–0.92	Backup

Results and impact

Speed, validation, and business outcomes

Speed vs. Traditional CRO Approach

Metric	Traditional CRO	GeneForge	Improvement
Timeline	8 to 10 weeks	10 business days	4× faster
Cost	$180K	$220K	Higher throughput, more candidates
Guides Evaluated	~50 manual designs	500K variants screened	10,000× larger search space
First-Pass Validation	Typically 60–70%	100% (5/5 guides)	No repeat iteration needed

In Vitro Validation Outcomes (3 weeks post-delivery)

T7E1 assay and amplicon sequencing results on top 5 guides.

Guide	Predicted On-Target	Observed Indel (%)	Off-Target Cuts	Notes
GF-g001	0.94	62%	0 detected	Highest cutting efficiency
GF-g002	0.92	58%	0 detected	Consistent across 3 donors
GF-g003	0.91	51%	0 detected	Best manufacturability score
GF-g004	0.89	47%	0 detected	Selected for backup series
GF-g005	0.88	44%	0 detected	Alternative splice site coverage

100%: First-pass validation rate (5/5)
0: Off-target cuts detected
62%: Max indel frequency (GF-g001)
3 wks: Timeline to validated guide

Immediate Wins

Manufacturing slot secured: top guide sequence submitted 6 weeks ahead of original CRO timeline
Preclinical budget preserved: no repeat design iteration required ($0 rework cost)
Investor diligence: presented genome-wide off-target analysis at Series A extension close

Strategic Advantages

Pathogenic loci exclusion list prevented guides that would have failed safety review
Liver-specific chromatin weighting improved hepatocyte on-target activity vs. generic scoring
Manufacturing-ready files eliminated translation gap between computational and CMC teams

Follow-on engagement
Q1 2026: AAV donor construct optimization for delivery vector. Estimated cost: $195K. Target: GLP-ready sequence package within 3 weeks.

Model validation

Lessons and recommendations

What Worked

Genome-wide off-target search at 500K scale caught 3 guides with cryptic pathogenic site matches missed by manual review
Liver chromatin accessibility weighting improved correlation with observed indel rates (R² = 0.91)
Manufacturability pre-filter prevented synthesis failures on 8 guides that scored well computationally

Challenges and Mitigations

GF-g003 had one low-risk off-target site in an intergenic region that required amplicon sequencing confirmation.

Mitigation: Added tiered off-target classification (critical/high/low/none) with explicit confirmation protocols per tier.

Two guides with high on-target scores showed reduced activity in one of three hepatocyte donor lines.

Mitigation: Added donor-line variability flag; recommended multi-donor confirmation for top 5 before manufacturing commitment.

When to use GeneForge for CRISPR guide design

In vivo gene therapy programs with strict off-target safety requirements
Teams without in-house computational biology capacity
Timeline pressure before manufacturing or IND-enabling studies
Rare disease targets with limited published guide data

ROI: approximately 4:1 (time savings + avoided repeat CRO iteration + manufacturing slot value).

Next steps: pair guide selection with vector optimization for delivery efficiency in target tissue.

About This Engagement

Client profile: Series A gene therapy startup, 8 employees, no computational team
Project duration: 10 business days (computational delivery) + 3 weeks (validation)
Total cost: $220K
Date: September 2025

This case study is anonymized at client request. Target gene, guide sequences, and institutional affiliations have been redacted. Full protocols available under NDA.

Run a similar program with GeneForge.

Tell us about your target. We will scope a 2 to 4 week pilot.

Schedule a Consultation [email protected]