🧬 biology

FINDER converts zero-background kinetic fingerprinting into area-scalable attomolar biomarker detection

The FINDER platform achieves area-scalable, amplification-free attomolar biomarker detection by combining fluorogenic transient probes with kinetic fingerprinting to suppress background noise and enable rapid single-molecule classification across large sensor areas.

Original authors: Nils Walter, Liuhan Dai, Pavel Banerjee, Alexander Johnson-Buck, Aaron Blanchard, Zi Li

Published 2026-06-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Nils Walter, Liuhan Dai, Pavel Banerjee, Alexander Johnson-Buck, Aaron Blanchard, Zi Li

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine you are trying to find a single, specific person in a massive, crowded stadium. This person is wearing a very specific, dull gray shirt. The problem? The stadium is filled with thousands of other people wearing similar gray shirts, and the lights are so bright that you can't see the difference between the real target and the crowd.

This is the challenge scientists face when trying to detect tiny amounts of disease markers (like specific strands of RNA or DNA) in a blood sample. Traditional methods try to spot these markers by how "bright" they are. But if the background is noisy, you either miss the real targets or count the wrong people.

Enter FINDER, a new technology developed by researchers at the University of Michigan and Duke University. Think of FINDER not as a camera that takes a snapshot, but as a security guard who watches people's behavior over time.

Here is how it works, using simple analogies:

1. The "Flashlight" Trick (Fluorogenic Probes)

Usually, if you shine a bright light (fluorescent probes) into a dark room to find someone, the light itself creates a glare that hides everything.

The Old Way: The "flashlights" are always on, even when they aren't touching the person you are looking for. This creates a sea of background noise.
The FINDER Way: The researchers used "smart flashlights" that are off until they actually hug the target. These are called fluorogenic probes. They are like glow-in-the-dark stickers that only light up when they stick to the specific person you are hunting. This means the room stays dark, and the only thing you see is the person you are looking for.

2. The "Handshake" Test (Kinetic Fingerprinting)

Even with the smart flashlights, some random people might bump into your target and stick for a second, creating a false alarm. How do you tell a real hug from a random bump?

The Old Way: You just look at how bright the person is.
The FINDER Way: You watch how they move.
- A real target has a specific "handshake" pattern. It grabs the probe, holds on for a specific amount of time, lets go, and then grabs it again. This happens over and over in a unique rhythm.
- A fake target (a false alarm) might just stick there forever or bounce off too quickly.
- FINDER watches these "handshakes" (binding and unbinding) for a few seconds. If the rhythm matches the secret code, it's a real target. If not, it's ignored.

3. The "Speed Run" (Why It's Fast)

In the past, watching these handshakes took a long time (minutes per spot), making it impossible to scan a large area.

The FINDER Breakthrough: Because the "smart flashlights" are so bright when they stick and so dark when they don't, the researchers can use a lot more of them (micromolar concentrations).
Imagine having a thousand security guards instead of one. They all rush to shake hands with the target at once. This makes the "handshake" happen much faster. Now, FINDER can verify a target in just 15 seconds instead of minutes.

4. The "Stadium Sweep" (Area Scalability)

This is the magic part.

The Problem: If you have a noisy background, scanning a bigger stadium (more area) just gives you more false alarms. You don't get better at finding the needle in the haystack; you just get more hay.
The FINDER Solution: Because FINDER has almost zero background noise (thanks to the smart flashlights and the handshake test), scanning a bigger area is like adding more searchers to a clean room.
- If you scan 10 spots, you find 10 times more targets.
- If you scan 100 spots, you find 100 times more targets.
- You don't get more false alarms.
- This allows them to find incredibly tiny amounts of disease markers—down to attomolar levels (that's one part in a billion billion).

What Did They Actually Find?

The paper reports that FINDER successfully:

Found a specific cancer-related RNA (hsa-miR-16) without needing to copy or amplify the sample first.
Detected a virus (HPV16) DNA.
Distinguished between two very similar DNA sequences that differ by only a single letter (like telling the difference between a "T" and a "C" in a genetic code), which is crucial for spotting cancer mutations.
Could look for RNA and DNA at the same time using two different colored "smart flashlights."

The Bottom Line

FINDER is like upgrading from a blurry, noisy security camera to a high-speed, smart system that watches how people behave. By combining "smart flashlights" that only turn on when they find the target with a "behavioral test" that checks the rhythm of the interaction, the system can scan huge areas quickly and find disease markers that were previously too rare to see without amplifying them.

Technical Summary: FINDER for Area-Scalable Attomolar Biomarker Detection

Problem Statement
Digital biomarker assays based on imaging, such as single-molecule nucleic acid counting, face a fundamental sensitivity constraint: the trade-off between surveying larger sensor areas to capture more analyte molecules and the accumulation of false positives. In conventional non-zero background detection, false positives accumulate linearly with the surveyed area (number of fields of view, $N$ ), causing the limit of detection (LOD) to improve only as $1/\sqrt{N}$ . While kinetic fingerprinting—classifying targets based on transient binding trajectories rather than static intensity—has shown promise in suppressing false positives, previous implementations required long observation windows (minutes per field of view) to establish definitive kinetic signatures. This throughput limitation prevented the practical sampling of large sensor areas, thereby blocking the realization of extreme sensitivity gains without enzymatic amplification.

Methodology
The authors introduce FINDER (Fluorogenic INstantaneous Digital Enumeration and Recognition), a platform designed to convert kinetic specificity into area-scalable sensitivity. The methodology relies on three core components:

Fluorogenic Readout Probes: FINDER utilizes transiently binding "imager" probes dual-labeled with a fluorophore and a quencher. These probes remain non-fluorescent in solution and only light up upon binding to the target. This design suppresses solution-phase background, allowing the use of micromolar probe concentrations ( $1\text{--}10\ \mu\text{M}$ ) to accelerate association rates without prohibitive optical noise.
Accelerated Kinetic Fingerprinting: By operating at high probe concentrations and tuning dissociation rates via engineered mismatches, temperature, and denaturants (e.g., formamide), FINDER generates multiple binding/dissociation transitions within seconds. This enables high-confidence classification of individual molecules in short observation windows ( $2\text{--}30$ seconds per field of view).
Multidimensional Kinetic Filtering: Instead of relying on intensity thresholds, the system classifies events based on kinetic features extracted from time-resolved trajectories, including the number of transitions ( $N_{\text{transition}}$ ), median dwell times ( $\tau_{\text{on,median}}, \tau_{\text{off,median}}$ ), and signal-to-noise ratio (SNR).
Background Suppression Strategies: To address surface-associated false positives (e.g., trapped probes), the authors employ a Quencher Probe Blocker (QPB) for miRNA assays and optimize surface passivation (PEGylation) and capture conditions for viral DNA.

Key Contributions and Results
The paper demonstrates that FINDER achieves a "near-zero-background" (ZBD) regime where false-positive counts remain constant as the surveyed area increases, while true target counts accumulate linearly. This shifts the sensitivity scaling from $1/\sqrt{N}$ to approximately $1/N$ .

hsa-miR-16 Detection: Using human microRNA-16 as a model, FINDER achieved an amplification-free LOD of 11 aM. The study validated that increasing the number of fields of view (from 1 to 40) with a 15-second acquisition time per field resulted in linear scaling of sensitivity. In contrast, shortening the acquisition to 1 second or using intensity-only thresholding resulted in higher LODs (695 aM and femtomolar range, respectively) and sublinear scaling due to background accumulation.
Viral DNA and Co-profiling: The platform was extended to detect Human Papillomavirus type 16 (HPV16) DNA, achieving near-zero blank acceptance with 30-second acquisition windows. Furthermore, a two-color assay successfully demonstrated simultaneous detection of HPV16 DNA and hsa-miR-16 RNA with minimal cross-channel interference, utilizing chromatic and kinetic feature spaces for separation.
SNV Discrimination: FINDER was applied to distinguish clinically relevant Single Nucleotide Variants (SNVs) in the EGFR gene (T790M and L858R mutations) from wild-type sequences. By optimizing formamide concentration and temperature, the system achieved specific discrimination within 2-second (T790M) and 10-second (L858R) observation windows per field, maintaining single-base specificity without amplification.

Significance and Claims
The authors claim that FINDER establishes a practical route to ultrahigh-sensitivity, amplification-free digital diagnostics. The central significance lies in operationalizing areal scaling: by accelerating molecular classification to the seconds-per-field regime while maintaining near-zero background, the platform allows sensitivity to scale linearly with the sensor area ( $1/N$ ). This converts the kinetic specificity of transient binding into a scalable sensitivity advantage.

The work generalizes the kinetic fingerprinting approach to diverse nucleic acid targets (miRNA, viral DNA, mutant DNA) and demonstrates the feasibility of multiplexed, amplification-free counting in practical workflows. The authors position FINDER as a generalizable framework for liquid biopsy and related applications, noting that while current workflows require empirical tuning of kinetic filters, the platform's architecture supports future automation and expansion to protein biomarkers. The paper explicitly states that these capabilities enable semi-automated attomolar biomarker counting without the need for enzymatic amplification.