FDA official Komudi Singh outlines draft guidance recommending NGS for genome‑editing safety assessments

Komudi Singh of the Food and Drug Administration’s Office of Therapeutics Products on Thursday summarized a newly published draft guidance advising how developers of human gene‑editing products should use next‑generation sequencing (NGS) in nonclinical safety studies. Singh urged stakeholders to review the document and submit comments to the docket where the draft guidance is posted.

The guidance, Singh said, builds on FDA’s January 2024 recommendations and focuses on nonclinical considerations for assessing off‑target editing and effects on genomic integrity. “This draft guidance provides recommendations for how developers of human gene therapy products incorporating human genome editing can use next generation sequencing technologies,” Singh said, describing the document’s scope and purpose.

Why it matters: Genome‑editing tools can produce edits at unintended genomic sites or create structural changes during DNA repair. Singh told attendees that off‑target edits that disrupt regulatory elements or coding sequences can harm cell function, and that erroneous repair after double‑strand breaks can cause chromosomal rearrangements. To detect low‑frequency or complex events, the draft recommends sensitive, quantitative NGS approaches and thorough reporting.

Key recommendations Singh highlighted include: using multiple nomination methods to identify candidate off‑target sites (biochemical assays, cell‑based assays, and computational in‑silico approaches where appropriate); documenting and validating any modifications to published assays; selecting NGS methods consistent with the editor’s mechanism of action; and choosing sequencing depth and input material sufficient to detect low‑frequency events. Singh emphasized that sponsors should document quality‑control metrics, analysis tools and steps, and record results for submission to FDA.

Reporting elements called for in the guidance, Singh said, include genomic coordinates, alignment information between the edit site and guide RNA (including mismatches or bulges), read counts and editing frequencies where applicable, annotation and use of published data to assess functional consequence, and gene‑expression information when it helps interpret functional impact. For in‑vivo products, he recommended using biodistribution data to determine which additional cell types should be assessed.

Singh also described how human genetic variation can influence off‑target risk: individual variants may make some genomic sites more similar to a guide RNA, increasing the likelihood of unintended editing. The draft guidance provides recommendations on database selection, variant filtering criteria and computational approaches, and asks sponsors to justify their chosen strategy.

For genome‑editing approaches that rely on double‑strand breaks, Singh said the agency recommends assessing chromosomal rearrangements at on‑target and confirmed off‑target sites with quantitative NGS methods and reporting coordinates, read counts or frequencies, annotation and a risk assessment.

Regulatory process and next steps: Singh encouraged sponsors to discuss study plans with FDA using INTERACT or pre‑IND meeting options and to include NGS safety study reports in IND submissions. He warned that additional analyses, raw sequencing data or analysis scripts may be requested during clinical or BLA review or after significant manufacturing changes.

The presentation closed with an invitation to submit public comments to the docket listed with the published draft guidance. The guidance is framed as recommendations to help sponsors plan nonclinical studies; it does not itself update statute or regulation, and Singh advised developers to consult the full document for details.