scTIM: seeking cell-type-indicative marker from single cell RNA-seq data by consensus optimization

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) provides high resolution for studying cell identity and its dynamics.
• Analyzing scRNA-seq data presents challenges due to noise, sparsity, and annotation difficulties.
• Identifying cell-type-specific marker genes is crucial for improving scRNA-seq data analysis.

Purpose of the Study:

• To develop a robust computational method for identifying cell-type-indicative markers from scRNA-seq data.
• To enhance the accuracy of cell type identification, annotation, and developmental trajectory reconstruction.
• To create a valuable tool for mining large-scale scRNA-seq datasets.

Main Methods:

• Developed scTIM, a method based on multi-objective optimization.
• Optimized for gene specificity, cell-cell relationship reconstruction, and gene redundancy reduction.
• Incorporated consensus optimization for enhanced robustness.

Main Results:

• scTIM demonstrated superior performance in cell type identification (clustering) and annotation across three diverse scRNA-seq datasets.
• The method effectively reconstructed cell development trajectories.
• Application to mouse cell atlas data generated a 'mouse cell marker atlas' for 15 tissues, revealing critical markers for tissue and cell type identities.

Conclusions:

• scTIM is an effective tool for identifying cell-type-indicative markers in scRNA-seq data.
• The method improves cell type annotation, clustering, and trajectory inference.
• scTIM provides a valuable resource for exploring tissue-specific cell types and their identities.