pipette
ESEspañol

Reframing decentralized diagnostics with a programmable multi-dimensional microfluidic system

Tao Xu, Ying Zhang, Xin Tang, Hongguo Wei, Lingyan Kang, Qize Zhou, Shunji Li, Xudong Zhao, Jiashuo Li, Zimeng Zhang, Chao Wan, Yufei Zhang, Dongjuan Chen, Ye Tian, Bi-Feng Liu, Peng Chen

Peer-reviewed journalBold claims, read criticallyReal-world use

In the authors' words

Decentralized diagnostics are vital for infectious disease control, yet current point-of-care tools remain largely one-dimensional, failing to capture the multidimensional profiles required for precision management. Here, we show SID-WAVES (Smartphone-Integrated Diagnostics with Wax-encoded Amplified and Versatile Evaluation System), an intelligent, multimodal microfluidic platform. Integrating low-melting-point phase-change and dual-siphon valves with RPA–CRISPR/Cas12a nucleic acid detection and protein immunoassays, the platform enables automated, one-step, sample-to-answer operation via an AI-assisted smartphone readout that reduces operator-dependent variability. Spatial-temporal-signal encoding supports multiplexed, multi-sample, and multimodal analysis, achieving attomolar-level sensitivity (10−18 M) and >95% diagnostic accuracy across 225 clinical and self-collected molecular assays, demonstrating robust usability. For proteins, a triple-amplification strategy and hierarchical capture enhance sensitivity 125-fold, improving accuracy from 87.6% to 100% in 54 simulated samples. Functioning as a reconfigurable analytical framework where panels are changed by swapping targets and probes, this work helps mitigate the Sensitivity–Multiplexing–Accessibility trade-off, reframing decentralized diagnostics from single-target screening to multi-dimensional health profiling. Decentralized diagnostics has the potential to improve health equity as well as support a more rapid response to disease outbreaks. Here authors present SID-WAVES, a multimodal microfluidic platform which allows ‘sample-in, results-out’ detection of both nucleic acid and protein biomarkers.

Main resultThe abstract does not state a limitation.

Appeared: Wednesday, October 7. Nature Communications. Peer-reviewed journal.

DOI: 10.1038/s41467-026-77744-1