An automated Bayesian pipeline for rapid analysis of single-molecule binding data

• Published in: Nature communications Vol. 10; no. 1; p. 272
• Main Authors: Smith, Carlas S., Jouravleva, Karina, Huisman, Maximiliaan, Jolly, Samson M., Zamore, Phillip D., Grunwald, David
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 17.01.2019; Nature Publishing Group; Nature Portfolio
• Subjects: 14; 14/63; 631/45/56; 631/57/2265; 9/10; Argonaute Proteins - metabolism; Automation; Bacterial Proteins - metabolism; Bayes Theorem; Bayesian analysis; Binding Sites; Computer simulation; Data processing; Deoxyribonucleic acid; DNA; DNA, Single-Stranded - metabolism; Endonuclease; Gaussian signals; Humanities and Social Sciences; Image Processing, Computer-Assisted - methods; Kinetics; Microscopy, Fluorescence - instrumentation; Microscopy, Fluorescence - methods; Molecular machines; multidisciplinary; Pipelines; Protein Binding; Proteins; Science; Science (multidisciplinary); Single Molecule Imaging - instrumentation; Single Molecule Imaging - methods; Software; State estimation; Thermus thermophilus - metabolism
• ISSN: 2041-1723; 2041-1723
• DOI: 10.1038/s41467-018-08045-5

Single-molecule binding assays enable the study of how molecular machines assemble and function. Current algorithms can identify and locate individual molecules, but require tedious manual validation of each spot. Moreover, no solution for high-throughput analysis of single-molecule binding data exists. Here, we describe an automated pipeline to analyze single-molecule data over a wide range of experimental conditions. In addition, our method enables state estimation on multivariate Gaussian signals. We validate our approach using simulated data, and benchmark the pipeline by measuring the binding properties of the well-studied, DNA-guided DNA endonuclease, TtAgo, an Argonaute protein from the Eubacterium Thermus thermophilus . We also use the pipeline to extend our understanding of TtAgo by measuring the protein’s binding kinetics at physiological temperatures and for target DNAs containing multiple, adjacent binding sites. Analysis of single-molecule binding assays still requires substantial manual user intervention. Here, the authors present a pipeline for rapid, automated analysis of co-localization single-molecule spectroscopy images, with a modular user interface that can be adjusted to a range of experimental conditions.