What if true variants could be distinguished from background noise with such precision that each individual variant had it’s own sensitivity threshold? Since noise varies across genomic positions, this kind of approach could be particularly useful in regions with high average noise, as the base locations with high noise could be identified and excluded from analysis. Yet most labs do not have this capability, as static sensitivity thresholds are often used in traditional limit of detection (LOD) measurements due to lack of sufficient samples to determine the LOD for each variant of interest.
That’s why scientists at Archer® recently developed a method to model the noise at each base position requiring only a small set of wild-type libraries. The position-specific noise profile is then used to calculate the LOD with per-variant resolution. Archer® Analysis expresses LOD as the minimum allele fraction at which a variant can be present in a sample and have a 95% probability of being detected above noise in a library, referred to as the 95 Minimum Detectable Allele Fraction (95 MDAF). Using this approach, position-specific noise profiling of TP53 exon 4 reveals that the high background noise is confined to only a few bases within the GC-rich region (shown below).