scTIM

scTIM identifies cell-type-indicative markers from single-cell RNA sequencing (RNA-seq) data to improve detection of cell identities and to reconstruct cell-cell relationships and developmental trajectories.

Key Features:

• Multi-Objective Optimization Framework: Employs a multi-objective optimization that simultaneously maximizes gene specificity via gene–cell relationships, maximizes genes' ability to reconstruct cell–cell relationships, and minimizes gene redundancy via gene–gene relationships.
• Consensus Optimization: Incorporates consensus optimization techniques to mitigate noise and variability for more robust marker identification.
• Clustering, Annotation, and Trajectory Reconstruction: Produces marker sets that improve clustering, cell-type annotation, and reconstruction of developmental trajectories from single-cell data.
• Validation on Diverse scRNA-seq Datasets: Validated on three diverse single-cell RNA-seq datasets to demonstrate robustness across biological contexts.
• Mouse Cell Marker Atlas Contribution: Identified critical markers across 15 mouse tissues contributing to the construction of a mouse cell marker atlas.

Scientific Applications:

• Cell Type Identification: Detects specific markers indicative of particular cell types to enhance characterization of cellular diversity.
• Developmental Biology: Enables reconstruction of cell development trajectories to investigate differentiation and maturation processes.
• Disease Research: Identifies subtle changes in cell populations that can inform studies of disease progression and cellular pathology.

Methodology:

Uses a multi-objective optimization framework optimizing gene–cell specificity, gene–gene redundancy, and genes' ability to reconstruct cell–cell relationships, combined with consensus optimization; validated on three single-cell RNA-seq datasets and applied to large-scale mouse cell atlas data covering 15 tissues.