Poster
Guide Profiler™: A Genetic Variant-Aware Computational Tool for Improved Guide RNA Selection for CRISPR-Based Therapeutic Applications
Summary
CRISPR-based gene editing has enormous potential for transforming genomic medicine. One of the challenges in realizing this potential is effective guide RNA (gRNA) design tools that go beyond on-target efficacy and incorporate specificity at potential off-target sequences influenced by population-specific genetic variation. Design tools not only need to move beyond the human reference genome to incorporate the genetic diversity we see in human populations but also robustly consider the relevance of potential off-target sites through meaningful biological annotation. Here we present Guide Profiler, which is a comprehensive approach to evaluate off-target risk during the guide selection process. This approach provides the ability to use in silico off-target enumeration and computational methods to rapidly screen large numbers of guide RNA sequences and to make an off-target informed list of viable guide RNA candidates. This will allow therapy makers to take into account ancestry of target donor and intent-to-treat populations, as well as to assess preparedness for diverse populations in clinical trial design. Off-targets are enumerated by performing a thorough sequence similarity-based search of the hg38 genome and 3,502 haplotype-phased human genomes from the 1000 genome and HGDP initiatives. Sequence search parameters are listed here:
| Parameter | |
|---|---|
| Maximum total differences from target sequence | 6 |
| Maximum mismatches (PAM and protospacer) | 6 |
| Maximum gap bases allowed (insertions, deletions) | 2 |
Off-targets are then annotated based on several biological factors including functional annotation, presence within and around protein-coding genes and clinical association to genes linked to cancer and Mendelian disease. To demonstrate use of Guide Profiler to perform guide selection based on off-target risk, results from eight guides each targeting the PCSK9 and B2M gene are presented here.
Figure 1: Guide Profiler overview (B2M)
A) Guide RNAs for B2M gene
B) Off-target enumeration
C) Impacts of SNV from South Asian population on on-target locus
D) Proximity of off-targets to on-targets
E) Off-target loci & ERC scores
F) Total number of loci for B2M guides
G) Guide rankings
Figure 1: Guide Profiler analysis was performed for guide RNAs targeting B2M gene. Example results are shown here. (A) Guide RNAs for B2M gene, exon 1 (CRISPOR software1). (B) Off-target enumeration with Levenshtein distance up to 6 using hg38 reference and 3502 globally diverse genomes. (C) Single-nucleotide variant (SNV) from the South Asian population impacts the on-target locus for B2M-1 and B2M-2. (D) Proximity of off-targets to on-targets – One off-target within 5 Mb was found for guides 4, 6 and 8. (E) Off-target loci are assigned an ERC score (Edits of Relative Concern) based on functional annotation. (F) Total number loci for B2M guides are displayed against a benchmark dataset of published guides. (G) Guide Profiler provides ranking of guide RNAs compared to other scoring methods.3,4,5 Guide Profiler takes into account several factors as described in Figure 2A.
Figure 2: Guide selection for PCSK9
In silico analysis using Guide Profiler was performed for eight guides targeting exon 1 of the PCSK9 gene. Available ONE-seq data for several guides (1, 3 & 4) supports in silico ranking (Figure 2B). This dataset bolsters the use of the Guide Profiler to perform an initial round of guide selection prior to collection of experimental data. Additional context for interpretation of off-target risk is provided through presentation of data on a set of published guides (Figures 1F, 2A).
A) PCSK9 guide scores & rankings
B) ONE-seq assay performance on 3 PCSK9 guides
C) Loci enumerated by Guide Profiler scores
Figure 2: (A) PCSK9 guides2 are scored and ranked by the Guide Profiler based on several criteria as shown. For benchmarking of PCSK9 guides, example data from published guides is shown. (B) ONE-seq assay was performed on three PCSK9 guides listed where there is concordance between Guide Profiler rank and number of nominated sites from the ONE-seq assay. Cancer genes that were observed to be highly edited in ONEseq assay are also listed here. (C) A weak relationship (R = 0.69) is observed between # loci enumerated and the Guide Profiler score, likely due to some locational randomness of off-target sites.
Figure 3: Biochemical data from ONE-seq® Screen
In silico ranking provided by the Guide Profiler provides rapid insight into a priori off-target risk for potentially dozens of candidate guide RNAs. This information can be used to eliminate high-risk guides and identify a subset of guides for which biochemical data is collected using an assay such as ONE-seq. Guide SelectTM in a multiplexed adaptation of ONEseq, allowing for efficient collection of information on several guides in a single run.
A)
B)
Figure 3: (A) Guide Select retains salient features of ONE-seq assay. (B) Limited sizes of enumerated off-target loci provides opportunity to multiplex guides in a single ONEseq run.
Conclusions
- The Guide Profiler provides in-depth off-target risk assessment, enabling informed guide RNA selection.
- Such analysis paves the way for collection of pertinent data to inform guide selection through use of scalable, multiplexed biochemical assays such as the ONE-seq assay (Guide Select).
- Integration of such tools in the discovery phase (versus at the IND-enabling stage) can de-risk a therapeutic program and help better inform decisions around moving a particular gRNA or set of gRNAs from discovery and into the clinic.
References
- Jean-Paul Concordet, Maximilian Haeussler. 2018. CRISPOR: intuitive guide selection for
CRISPR/Cas9 genome editing experiments and screens, Nucleic Acids Research, 46:W242–
W245. - Musunuru K,Chadwick AC, et al. 2021. In vivo CRISPR base editing of PCSK9 durably lowers
cholesterol in primates. Nature. 593(7859):429-434. - Hsu, P., Scott, D., Weinstein, J. et al. 2013 DNA targeting specificity of RNA-guided Cas9
nucleases. Nat Biotechnol. 2013. 31:827–832. - Doench JG, Fusi N. 2016. Optimized sgRNA design to maximize activity and minimize off-target effects of CRISPR-Cas9. Nat Biotechnol. 34(2):184-191.
- Labuhn, Maurice, Felix F Adams, et al. 2018. Refined sgRNA Efficacy Prediction Improves Large-and Small-Scale CRISPR–Cas9 Applications. Nucleic Acids Research. 46(3):1375–85.
