Trace DNA now represents over half of forensic submissions in Australia and often fails to produce reliable STR profiles due to degradation, inhibition, and low template levels. Despite advances in STR chemistry, current PCR workflows remain rigid. qPCR is used to assess sample quality, but amplification conditions are fixed, limiting recovery of usable genetic data from the most challenging evidence. 

We present a pioneering approach toward a fully adaptive "smart" PCR system that responds in real time to sample-specific challenges. Our strategy combines three innovations. First, a theoretical framework links PCR cycling parameters to STR profile quality. Second, a machine learning algorithm trained on experimentally derived profiles predicts optimal cycling conditions. Third, a novel integration of qPCR and STR amplification in a single reaction vessel generates both improved allelic peaks and real-time amplification feedback. 

Testing four commercial STR kits (GlobalFiler, Identifiler Plus, VeriFiler Plus, PowerPlex 21) with two qPCR systems (Investigator Quantiplex Pro, Quantifiler Trio) demonstrated full compatibility and allowed informative profiles to be produced while capturing real-time amplification curves. Under extreme inhibitory stress, where conventional STR kits fail, this combined system consistently yielded complete DNA profiles. Real-time data also enabled discrimination between inhibited and uninhibited samples early in the PCR cycling, establishing the feedback required for adaptive cycling conditions.  

Supplementing traditional PCR reactions with extra polymerase and buffer alone improved profile amplification of inhibited samples. Adding qPCR primers allows continuous monitoring and opens the pathway to dynamic adjustment of cycling conditions. Early implementation of a firefly-based optimisation algorithm improved allelic balance and reduced drop-in and drop-out events, confirming the predictive potential of a smart PCR approach. 

These results provide the first experimental validation of all core components required for a smart PCR system: monitoring, interpretation, and dynamic adjustment of amplification. While a fully autonomous instrument remains in development, previously challenging DNA samples can now be reliably amplified and informative profiles consistently generated. This work redefines the possibilities for forensic DNA analysis and offers a transformative, intelligent approach that could revolutionise the handling of trace and inhibited samples in laboratories worldwide. 

Authors

• Caitlin McDonald (Flinders University, Australia) 
• Duncan Taylor (Forensic Science SA & Flinders University, Australia) 
• Adrian Linacre (Flinders University, Australia)