scMTD

scMTD performs statistical multidimensional imputation of single-cell RNA sequencing (scRNA-seq) data to correct dropout events by integrating cell-level, gene-level, and transcriptome dynamics including pseudo-time.

Key Features:

• Multidimensional Imputation: Employs a statistical multidimensional approach that integrates cell-level, gene-level, and transcriptome dynamics to impute dropout-affected expression values in scRNA-seq data.
• Pseudo-Time Utilization: Incorporates pseudo-time information to capture temporal progression of cellular states for more informed imputation.
• Local Cell Neighbors and Gene Co-expression Networks: Identifies local cell neighbors and specific gene co-expression networks based on pseudo-time to guide imputation decisions.
• Performance Superiority: In real-data-based comparative analyses against state-of-the-art imputation methods, demonstrates improved FISH validation, trajectory inference, differential expression analysis, clustering, and cell type identification.
• Applicability to Various Data Types: Applicable to both Unique Molecular Identifier (UMI)-based and non-UMI-based scRNA-seq data.
• Enhanced Biological Signal Recovery: Recovers biological signals from transcriptomes to support analysis of gene expression dynamics and cellular heterogeneity.

Scientific Applications:

• Imputation of Dropout Events: Mitigates the impact of dropout events to provide more accurate expression matrices for downstream analyses.
• Study of Gene Expression Dynamics: Enables exploration of temporal changes in gene expression within individual cells using pseudo-time-informed imputation.
• Discovery of Rare Cell Types: Improves clustering and cell type identification to facilitate detection of rare cell populations.

Methodology:

Statistical model-based imputation integrating cell-level, gene-level, and transcriptome dynamics; use of pseudo-time to identify local cell neighbors and gene co-expression networks; evaluated in real-data comparative analyses including FISH validation, trajectory inference, differential expression, clustering, and cell type identification.