Quantifying sequencing error and effective sequencing depth of liquid biopsy NGS with UMI error correction

Liquid biopsies are a minimally invasive method to diagnose and longitudinally monitor tumor mutations in patients when tissue biopsies are difficult (e.g., in lung cancer). The percentage of cell-free tumor DNA in blood plasma ranges from more than 65% to 0.1% or lower. To reliably diagnose tumor mutations at 0.1%, there are two options: unrealistically large volumes of patient blood or library preparation and sequencing depth optimized to low-input DNA. Here, we assess two library preparation methods and analysis workflows to determine feasibility and reliability based on standards with known allelic frequency (0 and 0.13% in PIK3CA). However, the implementation for patients is still costly and requires elaborate setups. [GRAPHICS] METHOD SUMMARY Two widely used Illumina library preparation kits were benchmarked for next-generation sequencing of cell-free tumor DNA: one kit without unique molecular identifiers (UMIs) but in technical duplicates and one kit with UMIs. Two reference cell-free DNA samples were used with 0 and 0.13% tumor allele frequency, respectively. Targeted sequencing was performed to 50,000x average depth. Illumina's UMI Error Correction Local App was used for aligning and collapsing UMI sequences. Signal noise in UMI- and non-UMI libraries was compared, and effective sequencing depth loss due to the bioinformatic processing was quantified to allow for estimating the required sequencing depth.