Quality Control and Pre-Qualification of NGS Libraries Made from Clinical Samples

摘要

We have compared sequencing metrics from different types of clinical samples and different methods of making NGS libraries, for purposes of quality control of the samples and sample preps. We have performed the metrics using the HiSeq and MiSeq, however the same QC metrics could be measured on other platforms. By choosing metrics that can be measured from small amounts of data (e.g., 300,000 reads), we can measure the quality of the clinical samples and NGS libraries in a highly multiplexed format, before spending considerable time and money for downstream processes such as sequence enrichment and NGS. More predictive than measurement of insert size and concentration, these metrics predict whether the amount and quality of genomic DNA, as well as the sample preparation method is sufficiently efficient to generate high coverage from deep sequencing. Sequencing results from sonicated human and bacterial DNA, as well as human maternal plasma show that different sample prep methods yield libraries of very different diversity, uniformity of coverage and background. Library quality can also vary considerably in different lots of reagents and also different number of amplification cycles.By optimizing the efficiency of making genomic DNA into sequencing libraries fewer reads are necessary to achieve reproducible, high quality results, especially from limiting amounts of plasma, FFPE tissue, chromatin immune precipitates, and single cells.