Enhancing Single-Cell Trajectory Inference and Microbial Data Intelligence

The utilization of single-cell trajectory inference methods to deduce cell differentiation trajectories from single-cell transcriptomic or proteomic data holds significant importance in comprehending the developmental mechanisms of healthy tissues and offering valuable insights into pathological conditions. Nonetheless, the enhancement of accuracy and resilience in existing algorithms for inferring single-cell trajectories presents a persistent obstacle, mostly attributable to the interference caused by the identification of unrelated genes during single-cell sequencing. In order to effectively tackle this matter and expand the suitability of these methods to a wider array of biological data, we introduce iterTIPD, a trajectory inference method that utilizes iterative feature selection. The Iterative Topological Feature Selection (iterTIPD) algorithm is a widely employed approach in the field of genomics for the purpose of detecting differentially expressed genes. It is specifically designed to be applied repeatedly on linear or branching single-cell RNA sequencing data. The described iterative approach involves the selection of a subset of genes that make a significant contribution to the construction of the differentiation trajectory. This selection leads to enhanced precision and robustness in the ordering of cell pseudo-time. Furthermore, iterTIPD exhibits its efficacy not only in conventional single-cell data analysis but also in the realm of microbial data intelligence. The experimental findings obtained from the study of four scRNA-seq data sets demonstrate that iterTIPD significantly improves the accuracy and reliability of single-cell trajectory inference techniques. This enhancement renders iterTIPD a valuable resource for researchers across many fields, including the analysis of microbial data. Additionally, iterTIPD not only improves the efficiency of trajectory inference methods but also has robust generalization abilities. The iterTIPD method effectively reconstructs the differentiation track of neural stem cells, demonstrating a strong alignment with established brain progenitor cell differentiation pathways. Moreover, the present research demonstrates that Top2a and Gjal may serve as promising novel markers for characterizing activated neural progenitor cell subgroups. This finding underscores the algorithm’s capacity to uncover biologically significant information.