Xu Ke, Miglietta Luca, Kesakomol Piyanate, Holmes Alison, Georgiou Pantelis, Moser Nicolas, Rodriguez-Manzano Jesus
Department of Infectious Disease, Faculty of Medicine, Imperial College London, London W12 0NN, U.K.
Department of Electrical and Electronic Engineering, Faculty of Engineering, Imperial College London, London W12 0NN, U.K.
Anal Chem. 2025 Jul 15;97(27):14311-14320. doi: 10.1021/acs.analchem.5c01181. Epub 2025 Jul 2.
Multiplex PCR plays a critical role in diagnostics by enabling the detection of multiple targets in a single reaction. However, its use is often limited by the need for multiple fluorescent channels, which are restricted in standard PCR instrumentation. Amplification curve analysis (ACA) is a data-driven multiplexing (DDM) approach that overcomes this limitation by using real-time PCR data and machine learning to differentiate targets in a single-channel, single-well format, without requiring instrument modifications. As part of this DDM strategy, we previously introduced Smart-Plexer 1.0, a tool that simulates multiplex assays using empirical singleplex data to identify optimal assay combinations in silico, maximizing kinetic feature distances between targets to support ACA-based discrimination. While Smart-Plexer 1.0 performs reliably in controlled reactions and offers a strong framework for the ACA assay design, it relies on a single kinetic feature and a median-based distance metric, which limits its accuracy in reactions with variable target concentrations or efficiencies. Here, we present Smart-Plexer 2.0, a more robust and accurate version designed to improve the performance in amplification reactions affected by such variability. This version introduces three new kinetic features that are stable across different template concentrations and uses clustering-based distance measures to better capture the variability between targets. Compared to its predecessor, Smart-Plexer 2.0 reduces accuracy variance by an order of magnitude and improves ACA classification by 1.5 and 1% in retrospective 3-plex and 7-plex assays, respectively. In a multi-experiment, cross-concentration evaluation of a newly developed 7-plex assay, it achieved 97.6% ACA accuracy, confirming its robustness across complex scenarios. Smart-Plexer 2.0 offers a reliable and scalable way to design high-performance multiplex PCR assays using standard real-time PCR instruments.
多重聚合酶链反应（Multiplex PCR）通过在单一反应中检测多个靶标，在诊断中发挥着关键作用。然而，其应用常常受到对多个荧光通道需求的限制，而这在标准PCR仪器中是受限的。扩增曲线分析（Amplification Curve Analysis，ACA）是一种数据驱动的多重分析（Data-Driven Multiplexing，DDM）方法，它通过使用实时PCR数据和机器学习，在单通道、单孔格式中区分靶标，从而克服了这一限制，无需对仪器进行修改。作为这种DDM策略的一部分，我们之前推出了Smart-Plexer 1.0，这是一种利用经验性单重数据模拟多重分析的工具，以在计算机上识别最佳分析组合，最大化靶标之间的动力学特征距离，以支持基于ACA的区分。虽然Smart-Plexer 1.0在受控反应中表现可靠，并为ACA分析设计提供了强大的框架，但它依赖于单一的动力学特征和基于中位数的距离度量，这限制了其在靶标浓度或效率可变的反应中的准确性。在此，我们展示了Smart-Plexer 2.0，这是一个更强大、更准确的版本，旨在提高在受此类变异性影响的扩增反应中的性能。该版本引入了三个在不同模板浓度下都稳定的新动力学特征，并使用基于聚类的距离度量来更好地捕捉靶标之间的变异性。与其前身相比，Smart-Plexer 2.0将准确性方差降低了一个数量级，并在回顾性3重和7重分析中分别将ACA分类提高了1.5%和1%。在对新开发的7重分析进行的多实验、跨浓度评估中，它实现了97.6%的ACA准确性，证实了其在复杂情况下的稳健性。Smart-Plexer 2.0提供了一种可靠且可扩展的方法，可使用标准实时PCR仪器设计高性能的多重PCR分析。
