Rolling Signal-based Ripley's K: A new algorithm to identify spatial patterns in histological specimens

Abstract

The spatial distribution of cells within a tissue underlies organ function. However, these spatial distributions are often difficult to identify, making it challenging to evaluate how cells establish these patterns during development or how diseases may disrupt these patterns and impair function. To address this, we developed an image analysis tool based on a novel algorithm that identifies spatial patterns within tissues. This analytical tool was used to study the bone marrow, a specialized microenvironment in which spatial patterning of regulatory cells may influence the differentiation and survival of hematopoietic stem cells. Using this algorithm, we discovered clusters of regulatory cells within the bone marrow that suggest an organization of micro-niches, which may form the basis of the hematopoietic stem cell microenvironment. This work provides a new tool for the detection and analysis of tissue morphology that enables identification of spatial patterns within tissues that can lead to a deeper understanding of tissue function, provide clues for early onset of disease, and be used as a tool for studying the impact of pharmaceutics on tissue development and regeneration. In BriefThis work introduces a new statistic to analyze the patterning of cells and physiological features in histological images. This statistic was used on a published set of immunofluorescent images of murine bone to identify novel spatial structures within the bone marrow that may provide new inisghts to the organization of the hematopoietic stem cell microenvironment. HighlightsO_LIRSRK, a statistical tool for analyzing the spatial distribution of features in histological images, is introduced. C_LIO_LIRSRK incorporates the quantification of signal distribution to identify unique spatial patterns. C_LIO_LISpatial patterns in hematopoietic stem cell microenvironments are identified. C_LI Graphical Abstract O_FIG_DISPLAY_L [Figure 1] M_FIG_DISPLAY C_FIG_DISPLAY