Optimizing polymerase chain reaction (PCR) using machine learning

Abstract

Despite substantial standardization, polymerase chain reaction (PCR) experiments frequently fail. Troubleshooting failed PCRs can be costly in both time and money. Using a crowdsourced data set spanning 290 real PCRs from six active research laboratories, we investigate the degree to which PCR success rates can be improved by machine learning. While human designed PCRs succeed at a rate of 55-64%, we find that a machine learning model can accurately predict reaction outcome 81% of the time. We validate this level of improvement by then using the model to guide the design and predict the outcome of 39 new PCR experiments. In addition to improving outcomes, the model identifies 15 features of PCRs that researchers did not optimize well compared to the learned model. These results suggest that PCR success rates can easily be improved by 17-26%, potentially saving millions of dollars and thousands of hours of researcher time each year across the scientific community. Other common laboratory methods may benefit from similar data-driven optimization effort.

eDNAssay: A machine learning tool that accurately predicts qPCR cross‐amplification

“…A few studies have used machine learning to predict amplification in particular (and specificity, by extension). Models by Kayama et al (2021) and Cordaro et al (2021) had 70% and 81% accuracy, respectively. Better performance was achieved by Döring et al (2019), whose model produced an AUC of 0.953—slightly higher than our primer‐only model, but lower than our full‐assay model.…”

“…A few studies have used machine learning to predict amplification in particular (and specificity, by extension). Models by Kayama et al (2021) and Cordaro et al (2021) had 70% and 81% accuracy, respectively.…”

“…These algorithms work by associating patterns among user‐defined predictor variables with known outcomes, then model performance is assessed either internally (e.g., cross‐validation) or using a separate test data set. This modelling framework has quickly gained traction in molecular ecology (Fountain‐Jones et al, 2021), including eDNA‐based ecological assessments (e.g., Cordier et al, 2018) and primer design (e.g., Cordaro et al, 2021; Döring et al, 2019; Kayama et al, 2021). However, the ability of machine learning to assess qPCR assay specificity, particularly when hydrolysis probes are involved, has yet to be demonstrated.…”