使用人工构建的多重参考物质评估各种商用提取试剂盒从FFPE中提取的核酸质量与性能

Next generation sequencing (NGS) assays and other advanced genomic techniques for tumor profiling require sufficient and high-quality nucleic acid input. Formalin fixed and paraffin embedded (FFPE) samples are a common, yet challenging type of patient sample often used in these tests. Often in clinical labs, the patient specimen is limited, so pre-analytical handling is critical; maximizing nucleic acid yield from the sample is important, but quality and performance in downstream assays cannot be sacrificed. Furthermore, labs must consider the workflow of the extraction system, ease of use, and processing time. Highly consistent, whole process reference materials can greatly facilitate method evaluation. Here we assess the yield, quality, and performance of nucleic acids obtained with industry leading FFPE extraction kits using various Seraseq FFPE standards. Normal human cells were engineered to carry DNA and/or RNA variants. Variant allele frequency (VAF) and concentration in the cells were verified using the Bio-Rad QX-200 Droplet Digital PCR (ddPCR) System. The cells were formalin fixed, embedded in paraffin blocks and sectioned at 10-micron thickness. Extraction was performed with various commercially available kits from QIAGEN, Promega, Beckman Coulter, and AutoGen. Yields were analyzed using Qubit kits and ddPCR. TapeStation assays were used to assess fragment size and quality. VAF and concentration were again measured by ddPCR as well as several NGS assays. Yields, quality, and fragment size were variable between extraction kits. Interestingly, yields determined using Qubit did not always correlate with those measured by ddPCR. Similarly, for some DNA variants, there was a discrepancy between VAFs measured by different extraction kits, which seemed to be related to the GC content of nearby sequences. Increasing the salt concentration of buffer used in the extraction process rescued coverage of variants with lower nearby GC% regions indicating some extraction conditions can be detrimental to accurate variant quantification. Assessing performance of all steps of a tumor profiling workflow, from extraction to variant calling, is imperative to ensure high sensitivity and specificity. Extraction of nucleic acids from FFPE is a critical step that can affect the ability of the assay to accurately detect all variants in a patient sample. As shown here, a consistent standard that is readily available and contains many variants of varying complexity can be an invaluable tool in extraction assessment and optimization.