Lu Xiaoming, Pritko Daniel J, Abravanel Megan E, Huggins Jonah R, Ogunleye Oluwaferanmi, Biswas Tirthankar, Ashy Katia C, Woods Semaj K, Livingston Mariclaire W T, Blenner Mark A, Birtwistle Marc R
Department of Chemical and Biomolecular Engineering, Clemson University.
Department of Chemical and Biomolecular Engineering, University of Delaware.
bioRxiv. 2025 Mar 28:2024.10.23.619855. doi: 10.1101/2024.10.23.619855.
Genetically-encoded, single-cell barcodes are broadly useful for experimental tasks such as lineage tracing or genetic screens. For such applications, a barcode library would ideally have high diversity (many unique barcodes), non-destructive identification (repeated measurements in the same cells or population), and fast, inexpensive readout (many cells and conditions). Current nucleic acid barcoding methods generate high diversity but require destructive and slow/expensive readout, and current fluorescence barcoding methods are non-destructive, fast, and inexpensive to readout but lack high diversity. We recently proposed theory for how fluorescent protein combinations may generate a high-diversity barcode library with non-destructive, fast and inexpensive identification. Here, we present an initial experimental proof-of-concept by generating a library of ~150 barcodes from two-way combinations of 18 fluorescent proteins, 61 of which are tested experimentally. We use a pooled cloning strategy to generate a barcode library that is validated to contain every possible combination of the 18 fluorescent proteins. Experimental results using single mammalian cells and spectral flow cytometry demonstrate excellent classification performance of individual fluorescent proteins, with the exception of mTFP1, and of most evaluated barcodes, with many true positive rates >99%. The library is compatible with genetic screening for hundreds of genes (or gene pairs) and lineage tracing hundreds of clones. This work lays a foundation for greater diversity libraries (potentially ~10 and more) generated from hundreds of spectrally-resolvable tandem fluorescent protein probes.
基因编码的单细胞条形码在谱系追踪或基因筛选等实验任务中具有广泛用途。对于此类应用，条形码文库理想情况下应具有高多样性（许多独特的条形码）、非破坏性识别（在同一细胞或群体中进行重复测量）以及快速、廉价的读出方式（适用于许多细胞和条件）。目前的核酸条形码方法能产生高多样性，但需要破坏性且缓慢/昂贵的读出方式，而目前的荧光条形码方法是非破坏性的、读出速度快且成本低，但缺乏高多样性。我们最近提出了关于荧光蛋白组合如何生成具有非破坏性、快速且廉价识别功能的高多样性条形码文库的理论。在此，我们通过从18种荧光蛋白的双向组合中生成一个约150个条形码的文库，给出了初步的实验概念验证，其中61个经过了实验测试。我们使用混合克隆策略来生成一个经过验证包含18种荧光蛋白每种可能组合的条形码文库。使用单个哺乳动物细胞和光谱流式细胞术的实验结果表明，除了mTFP1外，单个荧光蛋白以及大多数评估的条形码都具有出色的分类性能，许多真阳性率>99%。该文库与数百个基因（或基因对）的基因筛选以及数百个克隆的谱系追踪兼容。这项工作为从数百种光谱可分辨的串联荧光蛋白探针生成更多样性的文库（可能多达10种及以上）奠定了基础。
