What Is The Most Important Metric For The Researcher When Choosing A Targeted Enrichment Solution – Alignment, Specificity, Or Uniformity?

The answer, quite simply, is none of the above. While all are important, alignment, specificity and uniformity are performance metrics that are secondary until one key question of performance has been addressed – coverage. Yes, coverage. And the first question is straightforward: Did the targeted selection method cover both known SNPs critical to the research, and important regions within the exon where unknown SNPs may exist?  This first question focuses on effectiveness. Simply put, did the method work where it needed to? Performance metrics that are marketed by manufacturers; however, focus more on efficiency and how well they work based on what the method covers. An interesting assessment, but this is much like putting the cart before the horse. Without the needed coverage, do these other performance metrics really matter?

Table 1 – BRCA Project Summary for TruSeq Custom Design

Number Of Targets	47
Number Of Attempted Amplicons	58
Number Of Gaps	15
Total Gap Distance	13172
Cumulative Target (BP)	19732
Redundancy	0
Coverage	33%

Let’s look at three BRCA 1 and 2 custom designs: TruSeq Custom Amplicon from Illumina, AmpliSeq from Life Technologies and TargetRich™ from Kailos Genetics. First, take a look at the results from the Illumina DesignStudio Team (Table 1). The Illumina project summary identified a cumulative target of 19,732 base pairs, which their design captured 6560. Coverage was 33 percent. Table One provides a more detailed summary. The results from AmpliSeq Custom Amplicon (Figure 1) were more similar to TargetRich (Figure 2) in cumulative target and base pairs covered. However, TargetRich, unlike the Life Technologies design, did not have any areas where coverage went to zero.

With any custom design, the selection of a target enrichment platform is likely driven by effectiveness…“How effective is the solution in covering my regions of research interest?” This begs the question: Are areas where the AmpliSeq design goes to zero important?  Furthermore, can additional sequencer capacity yield the needed read depth for a statistical call where the TargetRich design has lower coverage? Raising these questions, it is interesting that many manufacturers focus more on the efficiency of their solutions –alignment, specificity and uniformity – while often ignoring the most important metric, coverage.