Computational Molecular Biology 2025, Vol.15, No.2, 91-101 http://bioscipublisher.com/index.php/cmb 96 trajectory inference, asynchronously differentiated cells can be sorted to identify the starting point, end point and intermediate branch nodes of differentiation, thereby reconstructing the lineage relationship. Integrating the data of scRNA-seq and scATAC-seq can enhance the accuracy of trajectory inference. One approach is to combine the two types of data to construct a comprehensive low-dimensional embedding space, enabling trajectory inference to take into account both transcriptional and apparent information comprehensively. For example, the weighted Nearest neighbor (WNN) analysis introduced by Seurat v4 can integrate the adjacent information contributed by RNA and ATAC respectively to form a more reliable intercellular relationship map (Figure 2) (Lin et al., 2024). On this basis, trajectory calculation can avoid the pseudo-continuity that may occur solely based on transcription data. Especially in cases where some key transcriptional changes lag behind, ATAC data can provide leading indicators and improve the accuracy of trajectories. For instance, in the differentiation of immune T cells, the transcriptional changes of cell phenotypic transformation may occur after epigenetic changes. It might be difficult to sort with scRNA alone, but after the addition of scATAC, the early epigenetic changes correctly positioned the cells at the front end of the trajectory (Zhang et al., 2024). Figure 2 The architecture of scMI. (a) The overview of scMI. (b) A frequency-based RW algorithm with restart to sample subgraphs. The algorithm samples |$m$| subgraphs starting from the same node, and the final subgraph is obtained by filtering based on frequency. (c) Representation learning with inter-type attention heterogeneous graph neural networks. Graph convolutions preserve the topological structure information of the subgraph, while the inter-type attention mechanism aims to capture the implicit cross-modality relationships within the multi-omics data (Adopted from Lin et al., 2024) 6 Actual Case Analysis: Development of the Hematopoietic System or Nervous System 6.1 Integrated analysis of hematopoietic stem cell differentiation The hematopoietic system provides a classic model for studying the determination of cell fate. Hematopoietic stem cells (HSCS) are located at the top of the differentiation lineage and can generate all types of blood cells, including myeloid (red blood cells, granulocytes, megakaryocytes, etc.) and lymphoid (T cells, B cells, etc.). For a long time, the pathways and regulatory factors of hematopoietic differentiation have attracted much attention. However, there are still unsolved issues regarding the molecular mechanisms of cellular heterogeneity and fate determination at each stage of HSC differentiation. Important progress has been made in the integrated analysis of single-cell RNA-seq and ATAC-seq in this field (Lee et al., 2023).
RkJQdWJsaXNoZXIy MjQ4ODYzNA==