For as long as forensic scientists have run STR profiles through a capillary electrophoresis instrument, they’ve been arguing with a peak that isn’t real.
It shows up one repeat unit smaller than the true allele and it looks enough like a genuine minor contributor that an analyst has to stop and decide if it’s actual evidence. Multiply that decision by every locus in a mixture, every case in a backlog, and you start to understand why stutter has cost forensic labs more time than almost any other artifact in DNA typing.
A predictable problem in an unpredictable place
Stutter comes from strand slippage during PCR amplification. As DNA polymerase copies a tandem repeat region, the strands slip against each other and produce fragments one or more repeats shorter (occasionally longer) than the true allele. Analysts have known this for decades, and most labs have built careful thresholds and interpretation rules around it.
The trouble is that stutter shows up in exactly where a true low-level allele would also show up. That overlap creates what amounts to an interpretive paradox: the same peak position can mean either an artifact to discard or evidence to report, and getting it wrong runs in both directions. A stutter peak that clears the threshold becomes a false allele call. A true minor contributor that falls below the threshold never makes it into the profile.
In a single-source sample, this is a manageable nuisance. In a mixture, where peaks from multiple contributors already overlap and compete for attention, it becomes the difference between a profile that can be used to include or exclude a suspect and one that ends inconclusive.
Why this moment is different
Forensic labs have spent thirty years managing stutter downstream with filters, interpretation rules, and increasingly sophisticated probabilistic genotyping software. While those tools help, they don’t remove the underlying problem because the chemistry producing the artifact hasn’t changed.
That’s what makes the Paradyme™ 27GY STR System, currently in development at Promega, worth a closer look. It uses a reduced-stutter polymerase that lowers stutter peaks at the amplification step, rather than filtering them out afterward. It also pairs with our eight-color chemistry, which allows smaller amplicons across the loci most likely to suffer in degraded samples. Together, the system targets the two failure modes labs know best: false-positive alleles from elevated stutter and false-negative alleles from true minor contributors that get discounted as artifacts.
In our own early evaluation comparing a conventional STR chemistry to the Paradyme system on the same mock casework mixture, the difference at a single locus told the story. Where the conventional chemistry showed an artifact peak crossing the threshold and a true minor allele falling below it, the reduced-stutter chemistry showed neither problem. Cleaner peaks going into a probabilistic genotyping model mean fewer competing hypotheses for the software, and a more consistent experience for the analyst making the call.
The system also carries two built-in quality indicators, QIS and QIL, that give analysts more context about what’s happening in a sample before they start interpreting it. A comparison of the two ratios can point toward inhibition, degradation, or simply limited template. In practice, that’s the difference between sending a sample back for cleanup up front and spending significant time trying to force an answer out of a profile that won’t give a conclusive one.
What this actually changes
None of this erases the cost of adopting new chemistry. Validation, retraining, and accreditation management don’t pause for a better polymerase, and neither does the caseload sitting in the queue. That’s a legitimate reason to be cautious about any technology change, and forensic labs have every right to be discerning about it.
But the calculation shifts when the underlying problem finally has a real answer. A profile where the minor contributor is visible and typeable instead of buried under stutter is a case that reaches a conclusion rather than stalling out. Less time spent repeatedly reviewing borderline peaks is time that goes back into a lab’s actual throughput, without adding headcount. For lab directors answering to submitting agencies about turnaround, that’s worth the tradeoffs.
Thirty years is a long time to work around a limitation everyone in forensic DNA typing has learned to accept as the cost of doing business. The Paradyme™ 27GY STR System doesn’t make that history disappear, but it does mean the peak that isn’t there might finally stop taking up so much of an analyst’s attention.
Kari Siegenthaler
