6 December 2020 in bioRxiv
The advancement of behavioral analysis in neuroscience has been aided by the development of computational tools 1,2 . Specifically, computer vision algorithms have emerged as a powerful tool to elevate behavioral research 3,4 . Yet fully automatic analysis of social behavior remains challenging in two ways. First, existing tools to track and analyze behavior often focus on single animals, not multiple, interacting animals. Second, many available tools are not developed for novice users and require programming experience to run. Here, we unveil a computer vision pipeline called AlphaTracker, which requires minimal hardware requirements and produces reliable tracking of multiple unmarked animals. An easy-to-use user interface further enables manual inspection and curation of results. We demonstrate the practical, real-time advantages of AlphaTracker through the study of multiple, socially-interacting mice.
19 February 2021 in bioRxiv
Quantifying RNAs in their spatial context is crucial to understanding gene expression and regulation in complex tissues. In situ transcriptomic methods generate spatially resolved RNA profiles in intact tissues. However, there is a lack of a unified computational framework for integrative analysis of in situ transcriptomic data. Here, we present an unsupervised and annotation-free framework, termed ClusterMap, which incorporates physical proximity and gene identity of RNAs, formulates the task as a point pattern analysis problem, and thus defines biologically meaningful structures and groups. Specifically, ClusterMap precisely clusters RNAs into subcellular structures, cell bodies, and tissue regions in both two- and three-dimensional space, and consistently performs on diverse tissue types, including mouse brain, placenta, gut, and human cardiac organoids. We demonstrate ClusterMap to be broadly applicable to various in situ transcriptomic measurements to uncover gene expression patterns, cell-cell interactions, and tissue organization principles from high-dimensional transcriptomic images.
26 July 2025 in Bioinformatics
Motivation: Mutual nearest neighbors (MNN) is a widely used computational tool to perform batch correction for single-cell RNA-sequencing data. However, in applications such as spatial transcriptomics, it fails to take into account the 2D spatial information. Results: Here, we present spatialMNN, an algorithm that integrates multiple spatial transcriptomic samples and identifies spatial domains. Our approach begins by building a k-nearest neighbors (kNN) graph based on the spatial coordinates, prunes noisy edges, and identifies niches to act as anchor points for each sample. Next, we construct a MNN graph across the samples to identify similar niches. Finally, the spatialMNN graph can be partitioned using existing algorithms, such as the Louvain algorithm to predict spatial domains across the tissue samples. We demonstrate the performance of spatialMNN using large datasets, including one with N = 31 10x Genomics Visium samples. We also evaluate the computing performance of spatialMNN to other popular spatial clustering methods. Availability and implementation: Our software package is available on GitHub (https://github.com/Pixel-Dream/spatialMNN). The code is available on Zenodo (https://doi.org/10.5281/zenodo.15073963).
2 December 2025 in bioRxiv
The increased uptake of high-resolution spatially-resolved transcriptomics (SRT) technologies demands the development of unsupervised methods to extract cell types and their spatial distribution from biological tissues. However, unsupervised clustering is challenging due to the sparsity of the data and the differences in cell arrangement within tissues. Here, we introduce ClustSIGNAL, a spatial clustering method that adaptively uses neighbourhood information to overcome data sparsity and perform cell type clustering. ClustSIGNAL first defines initial clusters and subclusters of cells with similar gene expression patterns. For each cell, a fixed neighbourhood size is defined, and entropy is calculated based on the proportion of initial sub-clusters in the neighbourhood to capture its composition. Cell-specific weights, generated from entropy values, are used to embed spatial information into the gene expression through adaptive smoothing. The transformed gene expression is then used for clustering cell types. We compared our adaptive smoothing approach with other smoothing scenarios on four simulated datasets of varying spatial complexity. We also evaluated our clustering method on four publicly available high-resolution SRT datasets and compared its performance to that of three other spatial clustering methods. We showed that ClustSIGNAL performs multi-sample clustering with high accuracy and can identify subtle cell types and subtypes of biological relevance. It is also robust to changes in spatial structure of tissues, segmentation errors, and sparsity. Overall, ClustSIGNAL stabilises gene expression of cells in homogeneous neighbourhoods and preserves distinct gene expression of cells in heterogeneous regions, effectively balancing the use of neighbouring cells as prior knowledge for downstream analysis. The ClustSIGNAL R/Bioconductor package is available from bioconductor.org/packages/clustSIGNAL .