SPARK-seq:高スループットアプタマー探索および動態プロファイリングプラットフォーム(SPARK-seq: High-throughput Platform for Aptamer Discovery and Kinetic Profiling)

2026-01-05 中国科学院(CAS)

<関連情報>

• https://english.cas.cn/newsroom/research_news/chem/202601/t20260105_1145173.shtml
• https://www.science.org/doi/10.1126/science.adv6127

SPARK-seq: アプタマー探索と速度論的プロファイリングのためのハイスループットプラットフォーム SPARK-seq: A high-throughput platform for aptamer discovery and kinetic profiling

Guoyan Luo, Jia Song, Yongbo Fu, Yuanjie Jiang, […] , and Weihong Tan

Science  Published:1 Jan 2026

DOI:https://doi.org/10.1126/science.adv6127

Editor’s summary

In vitro selection can yield DNA or RNA sequences that form complex, three-dimensional structures and recognize specific target molecules. These aptamers are used in sensing, binding, and imaging applications, even for complex targets such as living cells, where the molecular species being recognized is not necessarily known. Luo et al. developed a sequencing method for aptamer target identification and a parallel computational model for aptamer prediction and design (see the Perspective by Mayer). CRISPR-mediated gene disruption provides perturbations that can be correlated to loss of binding, providing a signal to identify targets. The resulting high-throughput data provide a basis for generative machine learning to design optimized aptamer sequences. —Michael A. Funk

Structured Abstract

INTRODUCTION

Cell surface proteins represent the majority of clinically actionable targets and are vital to cellular communication, signaling, and homeostasis. However, methods for generating high-affinity molecular probes such as aptamers against these targets remain limited. Current approaches are low-throughput and often compromise native protein conformations, leaving most of these targets underexplored and hindering aptamer-based diagnostic and therapeutic development.

RATIONALE

We hypothesized that a multimodal approach combining CRISPR-based genetic perturbation with single-cell multiomics could overcome the limitations of conventional aptamer screening. By concurrently profiling genetic perturbations, gene expression, and protein binding within one single cell, we established single-cell perturbation-driven aptamer recognition and kinetics sequencing (SPARK-seq), a platform that enables high-throughput mapping of aptamer-target interactions in native cellular contexts. This integrated design not only facilitates the identification of binders to low-abundance targets but also directly couples discovery with kinetic profiling, accelerating the development of precision molecular tools.

RESULTS

We applied SPARK-seq to systematically interrogate aptamer-target interactions by combining multiplexed CRISPR knockout with single-cell mRNA and aptamer sequencing. Following four rounds of enrichment with Cell-SELEX (systematic evolution of ligands by exponential enrichment), the aptamer library was screened against a pooled population of cells subjected to CRISPR knockout of 13 surface proteins. Single-cell sequencing enabled simultaneous detection of gRNAs, transcriptomes, and aptamer binding events, linking genetic perturbations to aptamer-binding profiles.

Using our computational pipeline single-cell perturbation-aptamer recognition and targeted aptamer-generation algorithm (SPARTA), we analyzed 8466 high-quality single cells and clustered the top 10,000 unique aptamer sequences into 1906 families. Differential binding analysis identified 5535 aptamer sequences targeting eight distinct surface proteins. Orthogonal validation using flow cytometry, surface plasmon resonance, and microscale thermophoresis confirmed target specificity and nanomolar affinities. The approach also resolved complex binding patterns, including an integrin-targeting aptamer recognizing ITGA3/ITGB1 and a single aptamer binding multiple PTPR paralogs.

SPARK-seq preferentially enriched aptamers with slow dissociation rates (k_off). Across multiple targets, binding difference (-log₂FC) correlated strongly with k_off (R² = 0.8 to 0.9) but not with equilibrium affinity (K_D), underscoring the platform’s capacity to isolate aptamers with prolonged target engagement, a critical feature for therapeutic and diagnostic applications. Leveraging these data, SPARTA’s convolutional neural network–based classifier achieved approximately ~97% accuracy in predicting PTK7-binding sequences, and its generative module produced functional variants with improved kinetic profiles.

CONCLUSION

SPARK-seq is an integrated platform that unites CRISPR-mediated genetic perturbation, single-cell transcriptomics, and sequence-based aptamer profiling, augmented by the deep learning framework SPARTA. This work establishes a foundation for “Aptomics,” a multi-omics enabled framework for the large-scale, sequencing-driven decoding of aptamer binding specificities and target properties. By converting millions of aptamer-binding events into high-dimensional sequencing data anchored to genetic and transcriptional states, SPARK-seq enables direct genotype-phenotype-ligand mapping across thousands of individual cells. This multimodal strategy expands the dynamic range for detecting low-abundance and conformation-sensitive targets, resolves kinetic and enrichment patterns at sequence resolution, and supports rapid, scalable discovery and rational optimization of high-specificity aptamers, paving the way for advanced diagnostic and therapeutic applications.

Spark-seq for aptamer discovery and variant design.

A CRISPR-based Cell-SELEX workflow identifies aptamers by differential binding between control and knockout cells, followed by large-scale profiling of aptamer target interactions through single-cell sequencing. A deep learning framework then predicts and designs aptamer variants with improved binding kinetics.

Abstract

Cell surface proteins are key disease biomarkers and therapeutic targets, yet high-throughput methods for aptamer discovery targeting these proteins in situ remain limited. We introduce single-cell perturbation-driven aptamer recognition and kinetics sequencing (SPARK-seq), a high-throughput platform integrating single-cell messenger RNA and aptamer sequencing with CRISPR-based surface protein perturbation. In a single experiment, SPARK-seq simultaneously mapped 5535 distinct aptamers to eight surface proteins, capturing interactions across more than two orders of magnitude in protein abundance and spanning diverse biophysical classes. The method discriminated closely related paralogous proteins with no detectable cross-reactivity and provided kinetic information that enabled the prioritization of aptamers with slow dissociation rates. Leveraging this kinetic diversity, we engineered variants with improved off-rate properties. SPARK-seq establishes a platform for high-efficiency discovery and rational variant design of aptamers and functional nucleic acids, unlocking possibilities in diagnostics and therapeutics.