Contamination Assessment for Cancer Next-Generation Sequencing

Abstract

Context.— The presence of allogeneic contamination impacts clinical reporting in cancer next-generation sequencing specimens. Although consensus guidelines recommend the identification of contaminating DNA as a part of quality control, implementation of contamination assessment methods in clinical molecular diagnostic laboratories has not been reported in the literature. Objective.— To develop and implement a method to assess allogeneic contamination in clinical cancer next-generation sequencing specimens. Design.— We describe a method to detect contamination based on the evaluation of single-nucleotide polymorphic sites from tumor-only specimens. We validate this method and apply it to a large cohort of cancer sequencing specimens. Results.— Identification of specimen contamination was validated via in silico and in vitro mixtures, and reference range and reproducibility were established in a panel of normal specimens. The algorithm accurately detects an episode of systemic contamination due to reagent impurity. We prospectively applied this algorithm across 7571 clinical cancer specimens from a targeted next-generation sequencing panel, in which 262 specimens (3.5%) were predicted to be affected by greater than 5% contamination. Conclusions.— Allogeneic contamination can be inferred from intrinsic cancer next-generation sequencing data without paired normal sequencing. The adoption of this approach can be useful as a quality control measure for laboratories performing clinical next-generation sequencing.