Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Genome Editing Safety topic
No spam. Unsubscribe anytime.
FDA bioinformatics staff lays out NGS, multi‑method approach for off‑target assessment in genome‑editing INDs
Summary
FDA bioinformatics reviewers told developers to use a tiered approach—nominate, confirm and risk‑assess off‑target sites using multiple complementary methods, relevant human cell types, high‑sensitivity NGS, and to account for human genetic variation before IND submission.
Dr. Young Choy, a bioinformatics reviewer in FDA’s Office of Cellular Therapy and Human Tissues, presented agency expectations for off‑target assessment in genome‑editing products. He described three major safety concerns—off‑target editing, unintended on‑target outcomes, and chromosomal rearrangements—and recommended a tiered workflow: nomination of candidate off‑target sites, experimental confirmation, and a functional risk assessment.
Choy emphasized using multiple complementary nomination methods—computational in‑silico searches, cell‑based assays and biochemical assays—because each has different strengths (in‑silico is broad but assumption‑driven; cell‑based is biologically relevant but depends on delivery and repair; biochemical assays can be highly sensitive but may produce false positives). He said sponsors should justify the combination of methods and ensure that nomination is appropriate for the editor’s mechanism.
For confirmation, sponsors should perform targeted deep sequencing with adequate input material and depth to detect low‑frequency events, and report per‑site read counts, variant calls and editing frequencies. Documentation must include sequencing platform, library preps, bioinformatics tools and command‑line parameters, and quality metrics; raw data and analysis scripts may be requested during review. Choy urged using relevant human cell types (the target cell type for ex vivo products, representative tissues for in‑vivo delivery) and multiple donors to reflect chromatin context and genetic variability.
Because individual genomes contain millions of variants, he said in‑silico analyses should consider variant databases and that sponsors may need to test variant‑contributed candidate sites if common or medically relevant. For editors that create double‑strand breaks, sensitive assays for chromosomal rearrangements and translocations are necessary; any confirmed rearrangement should be tabulated with frequency and a functional risk assessment.
Choy advised that off‑target assessment should be completed before IND submission and discussed in INTERACT or pre‑IND meetings; draft guidance on NGS methods for off‑target assessment was noted as open for comment.
