Fig 2: Confocal images using red fluorescent silica nanoparticles.a–g Use fluorescent Si-NP and h–n uses fluorescent Si-NP' both added directly after sample incubation. All the images of membranes had anti-ApoAI as the capture antibody except for g and n that had Anti-ApoB. The images show the state of membrane for sample containing a 1 pM, b 10 pM, c 100 pM, d 1 nM and e 10 nM of PON1-HDL and h 1 pM, i 10 pM, j 100pM, k 1 nM and l 10 nM of HDL-P. A protein cocktail containing rePON1, reApoAI and HSA (1:40:1000) that contained an equivalent protein in 10 nM PON1-HDL and total HDL respectively, i.e., f 10 nM rePON1 + 400 nM reApoAI + 10000 nM HSA and m 1 nM rePON1 + 40 nM reApoAI + 1000 nM HSA were used as controls. g and n used 10 nM of HDL as sample but with anti-ApoB, thus providing no signal. Scale bars are 100\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mu$$\end{document}μm. Imaging was done once for each case.

Fig 3: Calibration of voltage shifts to target concentration.a–e Schematic showing the assay steps of the NGEMS platform. The washing step with 4×PBS and 0.1×PBS buffer is employed after each sample and nanoparticle incubation step before an CVC is recorded. a Sample is incubated for 20 min on the anti-ApoAI functionalized AEM surface and then the wash steps employed. b Captured HDL on the AEM surface that gives a CVC used as the baseline signal. c Addition of silica particles functionalized with anti-PON1 (Si-NP) gives a voltage shift depending on number of captured PON1-HDL. d Addition of silica particles with anti-ApoAI (Si-NP') binds to unoccupied HDL in c giving a shift proportional to PON1-free HDL. e Direct addition of silica reporters with anti-ApoAI (Si-NP') binds to all the HDL giving a shift proportional to HDL-P (total HDL). Silica is significantly larger than HDL thus allowing only one Silica nanoparticle per HDL. a–e and its key was created with biorender.com. f Typical CVC of two-step sequential (Si-NP and Si-NP') reporter addition strategy as shown in c and d for PON1-HDL and PON1-free HDL. g Typical CVC of one-step Si-NP' reporter addition strategy as shown in e for HDL-P (total HDL). Calibration plots for PON1-HDL h, PON1-free HDL i and HDL-P j shown as average of at least five replicates of each concentration with the error bars as the standard deviation. Same sample was measured repeatedly at every given concentration on different NGEMS sensors.

Fig 4: Control experiments using NGEMS.a Effect of delipidation on the CVC. Delipidation brings back the shifted CVC close to its baseline. Further addition of Si-NP/Si-NP' does not cause any shift but an addition of Si-NP" (silica attached to polyclonal anti-ApoAI) does cause a shift because of the presence of lipid-free ApoAI on the surface. Confocal image of the AEM surface b when treated with 100 pM HDL followed by addition of fluorescent Si-NP' and treated with PBS for 30 min as the control, c when treated with 100 pM HDL-P followed by addition of fluorescent Si-NP' and treated with PBS-Tween20 for 30 min for delipidation, d when predelipidated 100 pM PON1-HDL is used followed by addition of fluorescent Si-NP, and e when predelipidated 100 pM HDL-P is used followed by addition of fluorescent Si-NP'. For both d, e, detergent treatment is done at higher HDL concentration and then diluted with PBS to achieve desired concentration. f Control experiments with 50000× diluted pooled human and plasma protein cocktails at different orders of Si-NPs, and different capture antibodies and samples. Error bars represent one standard deviation and each case in f are triplicated with same sample measured repeatedly on different NGEMS sensors. Scale bars are 100\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mu$$\end{document}μm for b–e and were only repeated once.

Fig 5: Variation of signal with time and effect of mass-transfer limitation for a small sensor.a–d shows the signal at different times for 2.15 pM PON1-HDL, 215 pM PON1-HDL, 1.97 pM HDL-P, and 1000 pM HDL-P showing almost identical signals between 20 and 60 min. e Effect of antibody surface coverage on the overall signal with 100%, 5% and 0% coverage. Similar signal between 100% and 5% suggests we are in mass-transfer limited regime. f, g shows the concentration at the channel center along the channel and directly above the membrane (membrane located at origin), and how the concentration gradient decreases over time. Shared figure legend for both f and g shown in the latter. h Surface concentration of analyte on membrane surface and non-dimensional flux with time. i Zoomed-in snapshots of the microfluidic channel from numerical simulations in mass-transfer limited regime (irreversible reaction on the membrane surface). j Theoretical signal increase over time showing a pseudo-steady state after t* = 1. Each case in a–e was measured three times independently except for 20 min case of b, c where they were independently measured five times with error bars as one standard deviation. Same sample was measured repeatedly at every given concentration on different NGEMS sensors.