Engineering Scalable Multiplexed Molecular Screening Platforms

A central role of biomedical research is capturing and evaluating the rich information from genes, proteins, and/or other small molecules to provide actionable information. With modern advances and understanding, focus has shifted to sensing multiple analytes in parallel. Yet, further improvements in multiplexed tools are needed for improved throughput. Here we present two technologies that expand our toolbox for multiplexed molecular screening.Current standard in biochemical assay multiplexing follows the well-plate paradigm where microliter volumes are spatially separated and retained in wells, maintaining individual addressability of thousands of simultaneous reactions. Preservation of spatially indexed sample wells enables robotic liquid handlers to construct parallel combinatorial reactions. Droplet microfluidics offer facile means of translating reagents into nano- and picoliter reaction vessels at throughput speeds outpacing robotic liquid handlers. However, continuous droplet production usually requires serial methods of droplet fusion, mixing, and sorting. This loss in spatial indexing limits the ability to match well plates in screening combinations of different reagents or compounds. Demonstrated here is a microfluidic method coupled with phase-change gating technology to generate all pairwise combinations between two sets of reactants, enabling improved multiplexed screening of assay markers across multiple samples. This architecture can preserve spatial indexing with full on/off control of droplet merging, producing deterministic pairwise combinations of original droplets. Such an approach achieves a sharp 30-fold reduction in reagent consumption and nearly 40-fold reduction in liquid handling, suitable for a wide range of applications from medicine to agriculture.Understanding the molecular heterogeneity within the tumor microenvironment can help characterize host vs. neoplastic cell types, identify metastatic propensity, guide clinical management, or elucidate novel drug targets. Multiplexed molecular profiling that preserves spatial context traditionally requires either spectral differentiation, high-powered imaging equipment, or cyclic techniques involving serially degradative rounds of sample staining and antigen retrieval. Here we demonstrate an imaging platform, leveraging the phasor approach to fluorescence lifetime imaging microscopy (FLIM) and deep learning, to resolve up to 4 exogenous molecular probes. Resolution is achieved within a single spectral window, adding an orthogonal dimension with which to multiplex optical imaging