Choosing the Right Immunoassay Platform for Protein Quantitation
Fog City Bio R&D, LLC
Published 16 December, 2025
For Research Use Only (RUO)
Download the pdf version here.
Introduction
Selecting an immunoassay platform determines whether your experiment succeeds or fails before you pipette the first sample. Choose correctly and you generate clean, reproducible data. Choose incorrectly and you waste precious samples, burn through budget, and miss publication deadlines. The challenge is that there is no single “best” platform. The right choice depends on which constraint dominates your specific application.
Traditional guidance treats platform selection as a linear checklist: sensitivity, then matrix, then throughput. But researchers don’t approach the problem this way. A neuroscientist measuring femtogram-level biomarkers in cerebrospinal fluid faces entirely different constraints than a cell biologist profiling ten cytokines from limited sample volumes. The neuroscientist needs extreme sensitivity regardless of cost. The cell biologist needs multiplexing regardless of development complexity. Applying the same decision framework to both scenarios produces suboptimal recommendations.
This guide presents platform selection through the lens of limiting constraints. We organize the decision process around the factor that will kill your assay if you get it wrong: detection limits when you’re not sure you can see signal, or sample limitations when you have scarce material or need high throughput. A third section addresses troubleshooting scenarios when standard approaches fail. By identifying your primary constraint first, you can navigate directly to the platforms most likely to succeed rather than evaluating every option against criteria that may not matter for your application.
The platforms covered (ELISA, HTRF, AlphaLISA, Simoa, MSD electrochemiluminescence, and Luminex xMAP) span sensitivity ranges from nanograms to femtograms per milliliter and formats from single analyte to 500-plex. Understanding when each platform’s strengths align with your constraints prevents both under-engineering (choosing ELISA when you need femtogram sensitivity) and over-engineering (using Simoa for abundant proteins that ELISA detects easily). The goal is not to memorize specifications but to develop intuition for matching experimental requirements to technical capabilities.
Section 1: Sensitivity-Limited Decisions
When you don’t know whether you can detect your target protein, sensitivity becomes the primary constraint. This scenario occurs frequently with novel biomarkers, rare proteins, or attempts to measure established targets in challenging matrices. Literature may report the protein exists but provide no quantitative data. Pilot mass spectrometry might suggest low picogram concentrations but with insufficient precision to guide platform selection. In these cases, choosing a platform with inadequate sensitivity wastes irreplaceable samples and months of work.
Detection limits arise from the signal-to-noise ratio of the detection mechanism and the degree to which sample matrix contributes background. ELISA relies on enzymatic amplification: a horseradish peroxidase or alkaline phosphatase molecule conjugated to the detection antibody converts thousands of substrate molecules to colored product over the incubation period. Manufacturer specifications and optimized research protocols often cite detection limits of 10-100 pg/mL in optimal buffers, and exceptional cases with high-affinity monoclonal antibodies paired with chemiluminescent or fluorescent substrates can achieve limits as low as 1-5 pg/mL. However, these optimized limits represent best-case scenarios with extensive development effort. In routine practice, particularly when working with biological matrices, reliable ELISA performance typically begins around 50-100 pg/mL. The limiting factor is background signal from non-specific binding and matrix components. Serum proteins adhere to microtiter plate plastic despite blocking steps. Hemolysis contributes absorbance at common detection wavelengths. Heterophilic antibodies bridge capture and detection antibodies independent of target. These factors compress the usable detection range and often necessitate sample dilution that places low-abundance targets below the quantitation limit. The distinction between theoretical limits and practical performance is critical: an ELISA specification sheet claiming 10 pg/mL detection may be accurate for purified recombinant protein in buffer, but the same assay in plasma frequently requires 50-100 pg/mL concentrations for reproducible quantitation.
Time-resolved fluorescence platforms like HTRF address the background problem through temporal discrimination. HTRF pairs a long-lived europium cryptate donor fluorophore with a short-lived d2 or XL665 acceptor in a FRET configuration. When donor and acceptor are brought together by antibody-antigen binding, 337 nm excitation of europium triggers energy transfer and 665 nm emission from the acceptor. The critical innovation is measuring emission after a 50-400 microsecond delay. Biological autofluorescence decays within nanoseconds, but europium emission persists for milliseconds. By the time the detector opens, sample autofluorescence has vanished while specific signal remains strong. This temporal filtering typically achieves 10-100 pg/mL sensitivity in standard applications. With optimization (higher affinity antibodies, extended incubations, titrated reagent concentrations) HTRF can reach 1-10 pg/mL. The homogeneous format (no wash steps) further reduces variability since there are fewer opportunities for matrix components to interfere during processing.
AlphaLISA pushes sensitivity into the sub-picogram range through spatial proximity requirements rather than temporal discrimination. Donor beads generate singlet oxygen when excited at 680 nm. This reactive oxygen species has a four-microsecond lifetime in solution, during which it diffuses approximately 200 nanometers. If an acceptor bead resides within this radius, singlet oxygen triggers a cascade producing 615 nm emission. Random collisions between donor and acceptor beads rarely satisfy the proximity requirement long enough for signal generation. Only when antibody-antigen binding holds beads together does sustained signal occur. This mechanism provides exceptional specificity, rejecting background from matrix components that don’t form stable bridges between beads. Under optimal conditions with high-affinity antibody pairs, AlphaLISA routinely detects 1-100 pg/mL and can be pushed to 0.1-1 pg/mL. The limitation is that antibody affinity directly determines whether antigen binding can maintain the 200 nm proximity against thermal motion. Weak binding results in transient association and poor signal.
When targets fall below one picogram per milliliter, conventional analog immunoassays reach their physical limits. Simoa (Single Molecule Array) technology addresses this through digital counting rather than analog signal measurement. The assay captures enzyme-labeled immunocomplexes on paramagnetic beads using standard sandwich immunoassay chemistry, but then isolates individual beads in femtoliter-volume wells etched into fiber optic bundles. After sealing the array and adding fluorogenic substrate, each captured enzyme generates fluorescent product confined to roughly 50 femtoliters. This confinement produces locally high concentrations from single enzyme molecules (concentrations sufficient for detection by a CCD camera). The instrument images the entire array and counts how many wells contain signal (digital) rather than measuring total fluorescence (analog). Counting positive wells follows Poisson statistics, enabling quantitation down to 0.01-1 pg/mL (10-1000 fg/mL). Simoa has demonstrated reliable detection of prostate-specific antigen at 6 fg/mL in serum and neurofilament light chain at subfemtogram levels in cerebrospinal fluid. The platform’s limitation is reduced dynamic range compared to analog methods: once most wells contain beads with multiple enzymes, the digital-to-analog transition compresses the upper detection range.
Matrix complexity interacts with platform sensitivity in non-obvious ways. A protein at 50 pg/mL in buffer sits comfortably within ELISA’s theoretical detection range, but the same protein in neat plasma may produce unreliable quantitation or be undetectable. Plasma contains 60-80 mg/mL total protein, creating a million-fold excess of potential interfering molecules. High protein content increases non-specific binding to plastic and detection reagents. Endogenous proteases can degrade target or detection antibodies during incubation. Heterophilic antibodies in patient samples bridge capture and detection antibodies independent of target, generating false positives. These matrix effects often force 10- to 100-fold sample dilution to achieve acceptable background, which proportionally reduces target concentration. A 50 pg/mL protein diluted 1:10 becomes a 5 pg/mL detection problem well below ELISA’s practical limits. HTRF and AlphaLISA handle this scenario better not because they’re inherently more sensitive in buffer but because homogeneous formats and time/spatial discrimination mechanisms reject matrix interference more effectively, allowing work in more concentrated samples without dilution.
The decision framework for sensitivity-limited applications proceeds from target concentration estimate. For proteins expected above 100 pg/mL in clean matrices, ELISA provides adequate sensitivity at the lowest cost. Between 50-100 pg/mL, ELISA works in clean samples but becomes unreliable in complex matrices. HTRF becomes the safer choice for plasma or serum. In the 10-50 pg/mL range, HTRF is required for most applications as ELISA operates at its practical limit. In the 1-10 pg/mL range, optimized HTRF or AlphaLISA are necessary, as standard ELISA will fail. Below 1 pg/mL, only AlphaLISA (if highly optimized) or Simoa succeed. Below 0.1 pg/mL, Simoa is the only viable option. These ranges assume good antibody pairs and realistic development timelines; with poor antibodies or aggressive optimization requirements, performance degrades across all platforms and may require moving up one sensitivity tier.
References:
- Ren Y, et al. An Extremely Highly Sensitive ELISA in pg mL−1 Level Based on a Newly Produced Monoclonal Antibody for the Detection of Ochratoxin A in Food Samples. Toxins. 2023;15(8):493.
- Cheng CM, et al. A rapid and highly sensitive biomarker detection platform based on a temperature-responsive liposome-linked immunosorbent assay. Sci Rep. 2020;10:17974.
- Beaudet L, et al. AlphaLISA immunoassays: the no-wash alternative to ELISAs for research and drug discovery. Nat Methods. 2008;5:AN8-AN9.
- Xie B, et al. HTRF: A Technology Tailored for Drug Discovery. Curr Chem Genomics. 2009;3:9-20.
- Rissin DM, et al. Single-molecule enzyme-linked immunosorbent assay detects serum proteins at subfemtomolar concentrations. Nat Biotechnol. 2010;28(6):595-599.
Section 2: Sample-Limited Decisions
When sample volume is scarce, throughput demands are high, or multiple analytes must be measured from single aliquots, practical constraints dominate platform selection over pure sensitivity considerations. These scenarios shift the optimization target from “can we detect it?” to “can we execute the measurement plan with available material?” A pediatric clinical trial might collect 500 microliters of plasma per patient and need to measure twelve biomarkers. A high-throughput screening campaign might generate 10,000 samples requiring single-analyte quantitation within two weeks. Different constraints, different optimal platforms.
Sample volume limitations frequently arise in precious sample scenarios: pediatric studies, rare patient populations, needle biopsies, limited cerebrospinal fluid draws, archived specimens. When total available volume is 100 microliters or less and multiple measurements are required, assay volume becomes the dominant constraint. ELISA typically requires 50-100 microliters per well when run in duplicate, consuming the entire sample for a single analyte. Running five analytes would need 500 microliters which is impossible with a 100 microliter sample. Miniaturization to 384-well format reduces ELISA volume requirements to 20-30 microliters but introduces technical challenges with small-volume pipetting accuracy and increased evaporation during incubations. More fundamentally, running five separate ELISA assays, even in 384-well format, consumes more sample than a single multiplexed measurement.
Multiplexing addresses volume constraints by measuring multiple analytes from one sample aliquot. MSD electrochemiluminescence enables up to 10-plex measurements from 25 microliters through spatial encoding. Capture antibodies for different analytes are patterned as distinct spots within a single well of a specialized plate. Carbon electrodes beneath each spot provide the electrochemical trigger for light emission from ruthenium-labeled detection antibodies. When voltage is applied, only labels near the electrode surface (bound to captured antigen) undergo the oxidation-reduction reaction producing 620 nm emission. A CCD camera captures the entire well simultaneously, and software deconvolutes which signal corresponds to which analyte based on spatial position. This approach works well for cytokine panels where multiple analytes exist in similar concentration ranges and validated antibody sets are available. The limitation is that all analytes must be measured under the same assay conditions (buffer composition, incubation times, detection reagent concentrations). Targets with vastly different optimal conditions perform suboptimally in multiplex format.
Luminex xMAP technology extends multiplexing to 50-500 analytes through flow cytometric detection of color-coded microspheres. Each bead population carries a unique fluorescent signature created by mixing two internal dyes at different ratios. Beads are conjugated with capture antibodies specific for different analytes. After incubating with sample and phycoerythrin-labeled detection antibodies, the mixture flows through a dual-laser system. One laser interrogates internal dyes to identify which analyte the bead detects. A second laser quantifies surface-bound phycoerythrin, measuring how much antigen was captured. A single 25-50 microliter sample thus reports on dozens of analytes. The platform excels for pathway profiling, measuring comprehensive signaling cascades or all members of a cytokine family. The development burden increases with plex number since every antibody pair must be validated for cross-reactivity and optimal concentrations must be determined for each analyte within the shared assay buffer. For 2-5 analytes, the complexity may not justify multiplex development. Above 10 analytes, the sample savings become compelling.
Throughput considerations operate differently depending on whether you’re developing an assay or running production samples. During development, you’re running 10-50 samples to validate performance characteristics: linearity, precision, accuracy, sensitivity. Throughput is rarely limiting at this scale. Once you have a validated method and face 200-1000 samples for a study, throughput determines timeline and cost. ELISA becomes rate-limiting when sample counts exceed roughly 100. The multiple wash steps (typically 3-5 per assay) dominate hands-on time. A 96-well plate requires approximately three hours of labor: coating or sample addition, blocking, multiple incubation-wash cycles, substrate development, plate reading, and data analysis. Processing 500 samples in duplicate means six plates at three hours each, ~18 hours of hands-on work not counting setup and cleanup. For a single operator, that’s three full days of bench work.
Homogeneous assays like HTRF and AlphaLISA collapse this timeline by eliminating wash steps. After adding sample and pre-mixed detection reagents, the plate incubates and goes directly to the reader. A 96-well HTRF plate requires approximately one hour of hands-on time. The same 500-sample study needs six hours instead of 18 (3-fold reduction). This efficiency matters whether you’re running samples yourself or transferring a validated method to a client’s lab for large-scale production. HTRF and AlphaLISA also scale better to 384-well and 1536-well formats because the homogeneous format eliminates the complex liquid handling required for washing high-density plates. For screening applications processing thousands of samples, this scalability can determine whether a project is feasible.
Simoa occupies a unique position in the throughput discussion. The HD-X instrument processes up to 288 samples in an eight-hour day which is respectable throughput for a femtogram-sensitivity platform but slower than HTRF or ELISA. The SR-X benchtop instrument runs 96 samples in roughly three hours. However, Simoa’s value proposition isn’t throughput but rather enabling measurements that cannot be made any other way. If your target is at 100 fg/mL and invisible to every other platform, Simoa’s throughput is irrelevant and there is no alternative. The platform becomes the only option rather than an option compared against others on throughput metrics.
The decision framework for sample-limited applications asks different questions than the sensitivity-driven approach. If you have less than 50 microliters per sample and need multiple analytes, multiplex is essentially required: MSD for 3-10 targets, Luminex for 10+. If you have adequate volume but face high sample counts (200+), homogeneous formats (HTRF, AlphaLISA) reduce hands-on time substantially compared to ELISA. If sample count is modest (under 100) and only 1-2 analytes are needed, platform choice defaults back to sensitivity and matrix considerations since throughput is not limiting. Sample volume and throughput constraints can override sensitivity requirements: even if ELISA technically has adequate sensitivity, if you need five analytes from 75 microliters, multiplex becomes necessary despite higher cost.
References:
- Meso Scale Discovery. Electrochemiluminescence Technology White Paper. www.mesoscale.com
- Luminex Corporation. xMAP Technology Overview. Analyst. 2015;140(5):1174-1181.
Section 3: Troubleshooting and Edge Cases
Platform selection becomes complicated when standard decision frameworks produce unsatisfying answers or when initial attempts fail. These situations reveal the gap between theoretical platform capabilities and real-world assay performance. A protein should be detectable by ELISA based on literature concentrations, but your assay shows only noise. Alternatively, HTRF works beautifully in buffer but signal disappears in patient plasma. You need both femtogram sensitivity and 10-plex multiplexing, requirements that no single platform satisfies. Troubleshooting these scenarios requires understanding why platforms fail and when to abandon optimization attempts in favor of switching platforms entirely.
The most common failure mode is “buffer-to-matrix” collapse: an assay performs well in clean buffer but produces unacceptable background or loses signal when transitioning to biological samples. ELISA is particularly susceptible because the wash-based format provides multiple opportunities for matrix components to interfere. Serum albumin at 40 mg/mL can overwhelm blocking reagents and coat plate surfaces, creating non-specific binding sites. Heterophilic antibodies bridge capture and detection antibodies independent of target antigen, generating false signal. Lipemic samples scatter light, affecting absorbance measurements. The standard response is iterative optimization: try different blocking reagents (casein, fish gelatin, commercial blockers), vary wash stringency (buffer composition, number of washes, timing), dilute samples further. This optimization can consume weeks and substantial sample volume. After three rounds of optimization without improvement, switching to HTRF or AlphaLISA is often faster than continuing ELISA troubleshooting. The homogeneous formats simply reject more matrix interference by their fundamental mechanisms (time-gating for HTRF, and proximity requirements for AlphaLISA) rather than trying to wash interference away.
Signal irreproducibility despite acceptable sensitivity points to technical execution problems or antibody quality issues rather than platform inappropriateness. If coefficient of variation exceeds 20% between duplicates, the assay has fundamental reliability problems. Common causes include inadequate mixing (especially for bead-based assays where beads settle), temperature variation during incubations (uneven heating blocks), evaporation from edge wells of plates, inconsistent pipetting technique, and degraded reagents. These are platform-independent problems that switching technologies won’t solve. The diagnostic is whether problems persist across multiple proteins and antibody pairs. If yes, examine technique and equipment. If problems are target-specific, antibody quality is the likely culprit. Polyclonal antisera may work initially but lose titer over time as animals stop boosting. Monoclonal antibodies can denature with repeated freeze-thaw cycles. Before abandoning a platform, validate that antibodies still bind target by Western blot or direct binding assays. Many “platform failures” are actually reagent failures.
Conflicting requirements pose genuine technical dilemmas. A researcher needs to detect 0.5 pg/mL proteins (suggesting Simoa) from 50 microliter samples while measuring six analytes (suggesting MSD multiplex). Simoa doesn’t multiplex well beyond 2-4 analytes. MSD doesn’t reach subfemtogram sensitivity. No single platform satisfies both constraints. The pragmatic solutions involve compromise or sequential approaches. Option one: accept reduced sensitivity and use MSD 6-plex, acknowledging that the lowest-abundance analytes may be near detection limits. Option two: prioritize sensitivity and run six individual Simoa assays, accepting the sample volume consumption (300 microliters total) and increased cost. Option three: develop both a high-sensitivity Simoa assay for critical low-abundance targets and an MSD multiplex for the remaining analytes, running each on a different aliquot. The right choice depends on which constraint is negotiable. If detecting the rare protein is scientifically essential and the other five are secondary, Simoa becomes the priority. If comprehensive profiling matters more than ultimate sensitivity, MSD wins despite marginal detection of low-abundance targets.
Unusual matrices introduce problems not anticipated by standard validation protocols. Urine’s variable pH and salt content affects antibody binding. Saliva contains mucins that increase viscosity and interfere with pipetting. Fecal extracts contain pigments, proteases, and bacterial products that wreak havoc on immunoassays. Tissue homogenates release intracellular proteases and binding proteins that alter antigen availability. These matrices often require pre-treatment before immunoassay: pH adjustment, protease inhibition, lipid removal, size exclusion chromatography. The key question is whether pre-treatment is chemically compatible with the platform. HTRF’s europium cryptate labels are stable across pH 4-9 and tolerant of protease inhibitors. AlphaLISA’s beads aggregate in high-salt conditions but work after dialysis. Simoa is remarkably tolerant of matrix complexity due to multiple wash steps after bead capture but before array loading. This is because matrix components are washed away before detection, minimizing interference. When matrix pre-treatment is unavoidable, platforms requiring fewer washing steps during the assay (like HTRF) ironically become less advantageous since you’ve already committed to sample cleanup steps anyway.
Cross-reactivity issues compound in multiplex assays. An antibody pair perfectly specific for IL-6 in singleplex may show 2% cross-reactivity with IL-8. Negligible in isolation, but in a 10-plex cytokine panel where IL-8 is 50-fold more abundant than IL-6, that 2% cross-reactivity generates false IL-6 signal equal to the true IL-6 concentration. Extensive specificity testing against all panel members becomes essential for multiplex development but is often skipped in the rush to implement. The only solution is either accepting reduced specificity and carefully controlling for it through orthogonal measurements, or excluding cross-reactive analytes from the panel. We recommend rigorous cross-reactivity testing early in multiplex development since it’s faster to exclude a problematic analyte at the planning stage than to discover interference after running 300 patient samples.
When platforms repeatedly fail despite optimization, consider whether you’re measuring the right thing. Antibodies raised against recombinant proteins may not recognize endogenous proteins that undergo post-translational modifications. An ELISA for “total protein X” might miss a critical phosphorylated form that’s the actual disease biomarker. An immunoassay that works in one species may fail in another if epitopes aren’t conserved. Before concluding that ELISA lacks sensitivity, verify that your antibodies actually bind the endogenous form of your target. Spike recombinant protein into your matrix and confirm recovery. If spiked protein is detected but endogenous protein is not, your problem is antibody specificity, not platform sensitivity.
References:
- Han Z, et al. Relative Quantification of NaV1.1 Protein in Mouse Brains Using a Meso Scale Discovery-Electrochemiluminescence Method. Bio Protoc. 2021;11(3):e3910.
- Tighe PJ, et al. ELISA in the multiplex era: Potentials and pitfalls. Proteomics Clin Appl. 2015;9(3-4):406-422.
Conclusion
There is no universally “best” immunoassay platform. The optimal choice emerges from the intersection of technical requirements and practical constraints specific to each application. Researchers measuring abundant cytokines in cell culture supernatants face entirely different optimization problems than those detecting femtogram biomarkers in cerebrospinal fluid, and platform selection must reflect these differences.
The key insight is identifying which constraint dominates your specific scenario. When detection is uncertain (novel biomarkers, literature lacking quantitative data, or previous failed attempts) sensitivity becomes the primary filter. Target concentration drives platform choice: ELISA for targets above 50-100 pg/mL in clean samples, HTRF for the 5-100 pg/mL range particularly in complex matrices, AlphaLISA for 1-10 pg/mL applications, and Simoa when concentrations fall below one picogram. When sample limitations dominate with scarce volumes, high throughput demands, or multiple analytes from single aliquots, practical considerations override pure sensitivity optimization. Multiplexing becomes essential rather than optional when sample volume restricts how many times you can aliquot. Homogeneous formats (HTRF, AlphaLISA) dramatically reduce hands-on time for high-sample-count studies regardless of whether you need their full sensitivity.
Matrix complexity cuts across both sensitivity-limited and sample-limited scenarios as a modifying factor. Serum and plasma matrices attenuate effective sensitivity through interference mechanisms that increase background or require sample dilution. Platforms with built-in background rejection (such as time-resolved detection, proximity-based signal generation, or electrochemical excitation) handle complex matrices more robustly than ELISA’s wash-based approach. This explains why HTRF or AlphaLISA often succeed where ELISA fails even when theoretical sensitivity ranges overlap: matrix tolerance matters as much as raw detection limits.
Troubleshooting failures requires distinguishing platform limitations from execution problems or reagent issues. An assay working in buffer but failing in plasma may need platform change (ELISA to HTRF), but an assay with 30% CVs across duplicates has technical execution problems that platform switching won’t fix. Antibody quality issues masquerade as platform sensitivity problems until proven otherwise with recovery studies. Cross-reactivity in multiplex panels demands rigorous testing against all panel members before large-scale implementation.
For assay development services, understanding these decision frameworks allows efficient project scoping. When a client requests biomarker quantitation, the first questions reveal the dominant constraint: expected concentration, sample matrix, volume available, number of samples, number of analytes, timeline requirements. These answers immediately narrow platform options and set realistic expectations about development timeline and cost. Projects with conflicting requirements—ultra-sensitivity plus high multiplexing, for example—need transparent discussion about compromises or sequential approaches rather than promising impossible outcomes.
The practical reality for single-operator assay development laboratories is that we recommend platforms based on which will succeed reliably rather than which offers theoretical advantages. HTRF’s combination of good sensitivity, excellent matrix tolerance, and reasonable development time makes it our go-to recommendation for plasma biomarker quantitation in the 5-50 pg/mL range. AlphaLISA becomes the recommendation when pushing toward 1 pg/mL or when sample volume is severely limited. Simoa is the recommendation only when other platforms have failed or when literature clearly indicates femtogram concentrations. ELISA remains appropriate for targets above 50-100 pg/mL in clean matrices and for budget-constrained projects where the target is abundant. The goal is not to master every platform but to reliably match project requirements to platforms that will deliver results within realistic development timelines.
Disclaimer
This document is published by Fog City Bio R&D, LLC. It is intended for informational purposes only. All services and information provided by Fog City Bio R&D, LLC are strictly for Research Use Only (RUO) and are not intended for diagnostic or clinical use. Our services are not GxP-compliant and should not be used for in-human applications or regulatory submissions.
© Fog City Bio R&D, LLC — All rights reserved.