Genomics Evidence

 Quick Analysis Plan Copied

 Long Analysis Plan

What GUIDE-seq Directly Detects — and What It Structurally Cannot

GUIDE-seq tags double-strand breaks via dsODN integration, so its sensitivity is bounded by DSB formation, dsODN delivery efficiency, and sequencing depth; the original authors themselves flagged shallow sequencing and two-cell-line (U2OS, HEK293) reliance as sensitivity constraints (Original GUIDE-seq paper). A decisive stress test comes from spacer-nick Cas9n editing in HSPCs: off-target dsODN integrations were not detected at all in Cas9n-treated cells, and on-target GUIDE-seq reads dropped ~260-fold versus nuclease Cas9 (Spacer-nick off-target study). This is not proof of zero off-targets — single-strand nicks do not integrate dsODN — so GUIDE-seq under-reports nickase damage by design.

When GUIDE-seq-identified sites are then quantified by amplicon sequencing in edited cells, reported frequencies can be dramatic: in β-thalassemia base editing, the IVS2-1 guide produced off-target edits at S-OT2 (90.7% ± 2.3) and three E-OT sites above 50%, while the CD39 guide stayed mostly below 10% at its detected sites (β-thalassemia base editing off-target data). This spread shows why GUIDE-seq discovery must be paired with orthogonal quantitative validation — discovery sensitivity does not equal measured edit frequency.

Comparative Sensitivity: In Vitro Beats In Cellulo

CIRCLE-seq, a purely in vitro circularized-DNA assay, found 156 off-target sites missed by Digenome-seq with ~180,000-fold enrichment over background, illustrating that cell-based GUIDE-seq systematically misses chromatin-inaccessible or low-activity sites (CIRCLE-seq sensitivity paper). The trade-off: CIRCLE-seq false positives require cell-based confirmation — GUIDE-seq and CIRCLE-seq are complementary, not interchangeable.

Battle-Testing Verdict and Blind Spots

• Validated strength: GUIDE-seq correctly recovers known off-targets across nuclease, base-editor, and prime-editor systems (EMX1: 10 sites; FANCF: 3 sites in the SunTag-PE evaluation) (SunTag-PE GUIDE-seq assessment).
• Structural blind spot 1: nicks and SSBs are invisible by design (Cas9n data above).
• Structural blind spot 2: chromatin and cell-type dependence — K562 GUIDE-seq results may not transfer to HSPCs; the thalassemia study partially mitigates this with WES/RNA-seq orthogonal checks (Thalassemia study limitations section).
• Analysis-chain bias: off-target calling depends on pipeline assumptions (bulge handling, multi-mapper assignment); newer Snakemake pipelines address these but rely partly on in-silico validation (Modern GUIDE-seq analysis pipeline).

What would change this conclusion: a head-to-head study showing GUIDE-seq recovering nick-induced or low-frequency off-targets that CIRCLE-seq/amplicon validation miss would overturn the DSB-dependence blind spot; none exists in the supplied evidence. Confidence: high on reported values (all from supplied records), moderate on the generalization that GUIDE-seq under-reports genome-wide absolute off-target burden.