Advanced Imaging Center, Janelia Research Campus, Ashburn, Virginia, 20147.
Howard Hughes Medical Institute, Janelia Research Campus, Ashburn, Virginia, 20147.
Cytometry A. 2018 May;93(5):504-516. doi: 10.1002/cyto.a.23356. Epub 2018 Mar 13.
The spatial association between fluorescently tagged biomolecules in situ provides valuable insight into their biological relationship. Within the limits of diffraction, such association can be measured using either Pearson's Correlation Coefficient (PCC) or Spearman's Rank Coefficient (SRC), which are designed to measure linear and monotonic correlations, respectively. However, the relationship between real biological signals is often more complex than these measures assume, rendering their results difficult to interpret. Here, we have adapted methods from the field of information theory to measure the association between two probes' concentrations based on their statistical dependence. Our approach is mathematically more general than PCC or SRC, making no assumptions about the type of relationship between the probes. We show that when applied to biological images, our measures provide more intuitive results that are also more robust to outliers and the presence of multiple relationships than PCC or SRC. We also devise a display technique to highlight regions in the input images where the probes' association is higher versus lower. We expect that our methods will allow biologists to more accurately and robustly quantify and visualize the association between two probes in a pair of fluorescence images. © 2018 International Society for Advancement of Cytometry.
荧光标记生物分子在原位的空间关联为研究它们的生物学关系提供了有价值的见解。在衍射的限制范围内，可以使用 Pearson 相关系数 (PCC) 或 Spearman 秩相关系数 (SRC) 来测量这种关联，它们分别用于测量线性和单调相关。然而，真实生物信号之间的关系通常比这些措施所假设的更为复杂，这使得它们的结果难以解释。在这里，我们从信息论领域采用了方法，根据两个探针浓度的统计相关性来测量它们之间的关联。我们的方法在数学上比 PCC 或 SRC 更通用，不假设探针之间存在哪种关系。我们表明，当应用于生物图像时，我们的测量方法提供了更直观的结果，并且比 PCC 或 SRC 更能抵抗离群值和多种关系的存在。我们还设计了一种显示技术，突出输入图像中探针关联更高与更低的区域。我们预计，我们的方法将使生物学家能够更准确、更稳健地量化和可视化一对荧光图像中两个探针之间的关联。© 2018 国际细胞分析学会。
