How effective is cfDNA targeted sequencing panel for detecting low-frequency mutations?

Starting with 10 ng cfDNA is feasible, but stable detection of 0.1% VAF places very high demands on the workflow and data volume. Do not rely only on vendor examples; it is recommended to test dilution series of 0.5%, 0.1%, and 0.05% standards, and calculate deduplicated depth and reproducibility. Deduplicated depth is more critical than raw sequencing output; for 0.1%, at least 3,000x deduplicated depth is recommended, and the actual number of G required should be back-calculated based on panel size and on-target rate. High-GC regions are indeed prone to insufficient depth, which directly affects low-frequency mutation determination. High-GC exons of key genes should be separately assessed for coverage, and additional sequencing or increased data volume should be performed when necessary.

Buying as a bundle and buying separately each have pros and cons. For 10 ng low-input cfDNA, library preparation efficiency is highly sensitive. Bundled solutions typically have optimized workflows, with smoother after-sales and technical support, but the price is not necessarily lower. Buying separately offers flexibility, allowing you to use library preparation reagents you are familiar with, but you need to validate the workflow yourself, and the trial-and-error cost is not low. It is recommended to ask the vendor to quote both Panel-only and Panel plus library preparation complete solutions, calculate the cost per sample, and request data on low-input library preparation yield, coverage after deduplication, etc. If you do not have strong wet-lab experience, prioritize the bundled solution, which is at least more reliable for the first round of testing.

When selecting, focus on UMI design, on-target rate, off-target rate, coverage uniformity, and GC bias. Dual-end UMIs generally have an advantage in noise reduction for low-frequency mutations. You can have 2-3 suppliers perform small-sample testing using the same set of standards and real cfDNA samples, comparing the detection of variants around 0.1% and the depth after deduplication, rather than only looking at the data provided by the manufacturers. Operationally, pay attention to cross-contamination and index hopping, and try to use dual indexes. If certain high-GC regions have insufficient coverage and contain key genes, you can ask the manufacturer to supplement probes rather than accepting an overall decline in coverage. Ultimately, determine the sequencing data output based on the actual test results.