KSRV: a Kernel PCA-Based framework for inferring spatial RNA velocity at single-cell resolution.

Area of Science:

• Genomics
• Computational Biology
• Developmental Biology

Background:

• Understanding temporal gene expression dynamics in spatial contexts is crucial for deciphering cellular differentiation.
• RNA velocity analysis, by distinguishing spliced from unspliced mRNA, offers insights into future gene expression states.
• Current spatial transcriptomics technologies struggle to simultaneously capture spliced and unspliced transcripts at high resolution.

Purpose of the Study:

• To develop a novel computational framework, KSRV (Kernel PCA-based Spatial RNA Velocity), for accurate RNA velocity inference in spatially resolved tissues.
• To integrate single-cell RNA-seq with spatial transcriptomics data using Kernel Principal Component Analysis.
• To overcome limitations of existing spatial transcriptomics technologies in capturing both spliced and unspliced transcripts.

Main Methods:

• Developed KSRV, a computational framework integrating single-cell RNA-seq and spatial transcriptomics data.
• Utilized Kernel Principal Component Analysis for RNA velocity inference.
• Validated KSRV using 10x Visium and MERFISH datasets.

Main Results:

• KSRV accurately infers RNA velocity at single-cell resolution in spatially resolved tissues.
• Validation demonstrated KSRV's accuracy and robustness compared to existing methods like SIRV and spVelo.
• KSRV successfully revealed spatial differentiation trajectories in mouse brain and during mouse organogenesis.

Conclusions:

• KSRV is an accurate and robust computational tool for spatial RNA velocity analysis.
• The framework advances the understanding of spatially dynamic biological processes, particularly cellular differentiation.
• KSRV holds significant potential for future research in developmental biology and systems biology.