pearson correlation coefficients (r) calculated for pairwise combinations of relative gene quantities  (GraphPad Software Inc)

Journal: Frontiers in Neuroscience

Article Title: Analysis of clinical anatomical correlates of motor deficits in stroke by multivariate lesion inference based on game theory

doi: 10.3389/fnins.2025.1409107

Figure Lengend Snippet: Cross-correlations of lesion patterns across the positive contributors for the 272 patients. The strength of the Pearson correlation of pairs of regions of interest (ROIs) is color-coded from low (blue) to high (red). All correlations are statistically significant ( p < 0.05).

Article Snippet: We then computed pairwise Pearson correlation coefficients using MATLAB (Mathworks Inc., Natick, United States ) for the relative regional lesion patterns (i.e., correlations between all pairs of relative lesion sizes), across the 151 ROIs.

Techniques:

Journal: Cell Reports Methods

Article Title: A statistical approach for systematic identification of transition cells from scRNA-seq data

doi: 10.1016/j.crmeth.2024.100913

Figure Lengend Snippet: Identifying transition cells based on gene pairwise Pearson’s correlation coefficients (A) Stable cells and transition cells are separated through their intrinsic gene pairwise Pearson’s correlation coefficients (GPPCCs). In a Waddington’s landscape illustrating developmental processes, there are “valleys” and “ridges.” Valleys correspond to stable cellular states, and ridges represent barriers separating these stable states. During developmental processes, cells may transit from one stable state to another due to the change in the local landscape. We modeled these transitions as a result of the change in gene regulatory relations using stochastic differential equations (SDEs). Based on our mathematical derivations, gene pairwise correlation coefficients for transition cells are closer to ±1 compared with stable cells (illustrative heatmap: x axis, cells; y axis, gene pairs; color, values of GPPCCs) (see ). We further defined a transition index, which is proportional to the transition probability, to identify transition cells. (B) Transition cells identification workflow. To identify transition cells, we developed an analytical workflow containing several steps. We first did data preprocessing, including quality control, finding neighbors of each cell and obtaining the gene list with the largest expression variations. Then, GPPCCs were calculated for each cell by using the expression profiles of the cell and its nearest neighbors. Based on the empirical distribution of coefficients from all cells, a transition index, which is proportional to the transitioning probability, was calculated for each cell. (C)–(F) Identifying transition index using a simulation dataset. The simulation dataset is generated using SERGIO containing three steady states with linear transitioning structure. There are 5,000 stable cells in each steady state and 1,000 transition cells transitioning from state 1 to state 2 and state 2 to state 3. (C) UMAP of all cells with transition cells highlighted in red. (D) UMAP colored by transition index. (E and F) Evaluation with doublets. A total of 1,000 stable cells from state 1 and state 2 are randomly selected to generate doublets. (E) UMAP colored by the state of cells. (F) UMAP colored by transition index.

Article Snippet: This is because the strengths and connections of regulatory networks can change during cellular state transitions, , while CellTran requests gene pairwise Pearson’s correlation coefficients (GPPCCs) to be calculated from cells that have similar regulatory profiles.

Techniques: Control, Expressing, Generated

Journal: Cell Reports Methods

Article Title: A statistical approach for systematic identification of transition cells from scRNA-seq data

doi: 10.1016/j.crmeth.2024.100913

Figure Lengend Snippet: Transition index can accurately separate transition cells and stable cells (A) UMAP colored by cell types. Approximately 7,000 MuSCs from the mouse muscle regeneration dataset were used to validate the capability of CellTran to identify transition cells. Cell-type annotations are obtained from the original publication. (B) UMAP colored by transition index. Gray dots on the top right indicate inadequate observation of cells in the cluster to calculate transition indices. (C) eCDF of GPPCCs for transition cells (red) and stable cells (black) in the mouse muscle regeneration dataset. (D) Violin plot of transition index for stable cells, and transition cells in the mouse muscle regeneration dataset. Transition indices of transition cells are significantly higher than those of stable cells (Wilcoxon test; p < 0.01). (E and F) Performance comparison of CellTran, CellRank, and MuTrans in terms of (E) AUROC and (F) PRAUC using the mouse muscle regeneration dataset.

Article Snippet: This is because the strengths and connections of regulatory networks can change during cellular state transitions, , while CellTran requests gene pairwise Pearson’s correlation coefficients (GPPCCs) to be calculated from cells that have similar regulatory profiles.

Techniques: Comparison