BGPT: Paper Review: A highly specific CRISPR-Cas12j nuclease enables allele-specific genome editing

Fuel Your Discoveries

Quick Explanation Copied

Cas12j-8 (TTN PAM, mismatch-sensitive near PAM) is positioned as a compact CRISPR nuclease for allele-specific SNP disruption.

Functional screen: among 6 Cas12j orthologs, Cas12j-8 is reported as efficient and highly specific in human cells.
Specificity claim: mismatch tolerance is strongly reduced for PAM-proximal positions in crRNA designs.
Allele specificity claim: the paper reports SNP-discriminating disruption at 4 SNP loci with low off-target at non-target alleles.
Scope claim: in silico counting is used to estimate many ClinVar/dbSNP variants potentially targetable with allele-specific disruption using a TTN PAM.

Primary source: https://www.science.org/doi/10.1126/sciadv.abo6405

Long Explanation

Paper Review (BGPT): "A highly specific CRISPR-Cas12j nuclease enables allele-specific genome editing"

Publication: Sci. Adv., Feb 12, 2026 (DOI: 10.1126/sciadv.abo6405)

Core claim: a compact Cas12j ortholog (“Cas12j-8”) that is mismatch-sensitive near the TTN PAM enables allele-specific SNP disruption in human cells.

1) Evidence map: what the paper actually demonstrates

Ortholog functionality screen: the paper reports a GFP activation reporter system used to compare 6 Cas12j orthologs, identifying Cas12j-8 (and Cas12j-10) as producing GFP-positive edited cells, while others do not.
PAM specificity: the paper reports TTN preference for Cas12j-8 (and Cas12j-10) using deep sequencing readouts from PAM-library strategies.
Mismatch sensitivity / “specificity”: the paper uses GFP reporter panels with dinucleotide and single-nucleotide mismatches, plus an in vitro digestion assay, concluding Cas12j-8 is especially sensitive to mismatches at PAM-proximal positions.
Genome-wide off-target profiling (within scope): GUIDE-seq is reported for selected loci and crRNAs, with the paper claiming FnCas12a shows off-target sites while Cas12j-8 shows none at the AAVS1 site 1 locus and few across a limited set of targets.
Allele-specific SNP disruption (experimental): the paper reports allele-discriminating disruption at 4 SNP loci in human cells, with an emphasis on low off-target at non-target alleles.
Base editor attempts: the paper reports engineering a dead Cas12j-8 (E567A) and fusing it to an adenine base editor deaminase to create Cas12j-8ABE8e (938 aa), with reported activity at a subset of loci; it also reports a cytosine base-editor fusion with no detected editing at 50 loci in HEK293T.
In silico variant counting: the paper reports estimated targetability for allele-specific disruption—25,931 clinically relevant variants in ClinVar and 485,130,147 SNPs in dbSNP (and compares to Un1Cas12f1).

Skeptical reading

All “allele-specific” confidence in this paper ultimately hinges on (i) mismatch sensitivity near the PAM being reliably translated into allele discrimination in endogenous chromatin contexts, and (ii) the GUIDE-seq survey covering the relevant mismatch-driven off-target landscape. The paper does not claim exhaustive genome-wide safety across all cell types or in vivo immune/repair contexts—those remain unknown from the provided text.

2) Visual: TTN-allele disruption landscape (as claimed)

The paper provides point estimates from in silico analyses. Below, I visualize those counts (no inference beyond what is explicitly stated).

Critical note: These are counts, not experimentally validated editing outcomes. The paper’s therapeutic relevance requires additional validation for editing efficiency, mismatch discrimination under realistic delivery and repair conditions, and safety/off-target profiling beyond the limited GUIDE-seq scope.

3) Technical critique: what “high specificity” could mean here

3.1 Specificity evidence types used

Reporter specificity: GFP activation assays with mismatched crRNAs provide a functional readout correlated with in-frame indels.
Purified in vitro cleavage specificity: mismatch sensitivity is also probed in vitro.
GUIDE-seq off-target discovery: GUIDE-seq is used for genome-wide DSB profiling, but only across selected loci and conditions.
Allele discrimination: allele-specific editing at 4 SNP loci is used as the functional “discrimination” demonstration.

3.2 Key unknowns / potential failure modes

Mismatch sensitivity ≠ universal allele specificity: mismatch tolerance is position-dependent in their assays, but in real genome contexts, chromatin accessibility and local repair dynamics can change the mapping from PAM-proximal mismatch sensitivity to allele discrimination. (This is a logical gap; the paper’s text doesn’t provide full chromatin-state generalization.)
Guide selection / PAM selection bias: their allele-specific candidate list depends on the TTN PAM and SNP contexts they choose. The paper reports in silico counts, but we cannot infer that every counted variant behaves identically experimentally.
Off-target discovery sensitivity: GUIDE-seq depends on experimental conditions and detection limits; “no off-targets detected” is contingent on detection power and the selected loci/conditions tested.
AAV delivery and base editor performance: Cas12j-8 is claimed deliverable by AAV, but for Cas12j-8ABE8e the paper reports low detectable editing in their AAV attempt; for the cytosine editor fusion, no activity is reported across 50 loci in HEK293T.

4) Visual: base-editor edit scope (paper-reported)

The paper reports editing at 4 loci out of 20 tested for Cas12j-8ABE8e, and 0 out of 50 for Cas12j-8CBE-hA3A (in HEK293T).

Critical note: “Editing at a locus” is not equivalent to quantifying robust fraction of alleles edited, nor to demonstrating allele-specific discrimination for the base editor; the paper’s AAV delivery note implies efficiency limitations for the base-editor fusion.

5) Overall assessment

What looks strong

The paper triangulates specificity using (i) reporter mismatch panels, (ii) purified in vitro cleavage, and (iii) GUIDE-seq comparisons, then closes the loop with experimental allele-specific SNP disruption at multiple loci.
The TTN PAM simplicity is presented as a mechanistic driver for a high target density, supporting the in silico “coverage” narrative.

What remains uncertain / where results could mislead

In vivo translation is not established in the provided text. The paper includes AAV delivery demonstrations for indel editing, but key safety/efficacy properties for allele-specific correction/disruption in living organisms remain unproven here.
Allele-specific claims depend on mismatch-proximal effects. If mismatch effects shift under different delivery kinetics or repair pathway contexts, discrimination could degrade. The paper’s logic is sound, but the generality is not exhaustively tested across diverse loci/cell types.
Guide-count vs. experiment-count gap. The variant numbers are informative for design space, but they are not equivalent to proven allele-specific editing at each counted variant.

What would most change the conclusion?

Demonstrating that allele discrimination fails (i.e., substantial non-target allele disruption) at many additional SNP loci beyond the 4 tested, especially under delivery conditions that reproduce realistic editing kinetics.
Revealing off-target activity at additional targets or broader conditions where GUIDE-seq detection differs (cell type, chromatin state, dosage).
Showing that AAV-delivered Cas12j-8ABE8e (or other fused editors) can’t reach the needed editing fractions due to delivery/efficiency constraints.

Explore author-specific critical reviews on BGPT

Feedback:

Updated: April 03, 2026