bioRxiv · 10.64898/2026.11.12.766091
Population-aware computational assessment of base-editing accessibility among pathogenic variants on human chromosome 21: an African-ancestry-focused pilot study
Abstract
Background Population-aware gene-editing research requires independent evaluation of clinical variant evidence, ancestry-stratified allele observations, nucleotide correction chemistry, and sequence-level editing constraints. We developed a source-traceable chromosome 21 pilot connecting these dimensions without equating computational accessibility with therapeutic editability. Methods Reviewed ClinVar pathogenic/likely pathogenic single-nucleotide variants were matched by GRCh38 chromosome, position, REF and ALT to overall and African/African-American (AFR) gnomAD v4.1 exome allele counts. Exploratory frequency contrasts used two-sided Fisher exact tests and Benjamini-Hochberg correction. Higher-confidence variants with canonical adenine or cytosine base-editing chemistry underwent NGG PAM/editor-window screening, bystander enumeration, MANE v1.5 consequence annotation, theoretical joint-product assessment and chromosome 21-restricted guide-homology searches. Results Among 1,944 ClinVar variants, 851 matched gnomAD; 694 were quality-filtered and alternate-allele-observed overall, including 300 higher-confidence records. Six variants met exploratory AFR-elevation criteria. Of 234 higher-confidence ABE/CBE chemistry-compatible variants, 118 (50.4%) had at least one nominal NGG/PAM-window placement, 55 (23.5%) a nominal zero-bystander placement, and 44 (18.8%) a zero-bystander placement without an additional chromosome 21 NGG near-match within three mismatches. The 261 placements generated 295 hypothetical bystander events (195 distinct alleles); five distinct potential products overlapped pathogenic/likely pathogenic ClinVar alleles. Across placements, 764 theoretical target-corrected sequence states were enumerated, of which 503 contained at least one bystander. Three variant-level editing-accessibility comparisons between AFR-observed and AFR-unobserved alleles were not statistically significant (BH q=1.0). Conclusions Canonical correction chemistry substantially overestimates nominal local base-editor accessibility. A population-aware, transcript-informed computational framework can identify hypotheses for follow-up, but allele-frequency contrasts, guide-similarity results and enumerated sequence products do not establish clinical efficacy, experimental editing specificity or genome-wide off-target safety.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Achab-Awelana, E., Yeboah, E. K., Aryeh, C. T., Essandoh, A.. 2026-11-16. Population-aware computational assessment of base-editing accessibility among pathogenic variants on human chromosome 21: an African-ancestry-focused pilot study. https://doi.org/10.64898/2026.11.12.766091
Cite the original work for its findings. Save a collection to share your selection of sources.