TransST: Transfer Learning Embedded Spatial Factor Modeling of Spatial Transcriptomics Data.

Area of Science:

• Biomedical Research
• Genomics
• Computational Biology

Background:

• Spatial transcriptomics offers insights into tissue RNA profiles but suffers from low resolution and sequencing depth.
• Extracting reliable biological signals from spatial transcriptomics data remains challenging.
• Existing methods struggle to accurately differentiate cell types and identify specific tissue components.

Purpose of the Study:

• To develop a novel transfer learning framework, TransST, for spatial transcriptomics data analysis.
• To leverage external cell-labeled information to improve cell-level heterogeneity inference.
• To enhance the identification of cell subclusters and their driving biomarkers.

Main Methods:

• Proposed a novel transfer learning framework named TransST.
• Utilized adaptive leveraging of external cell-labeled information.
• Applied the framework to infer cell-level heterogeneity in target spatial transcriptomics data.

Main Results:

• TransST significantly improves upon existing techniques in both simulation and real-world studies.
• Successfully identified five biologically meaningful cell clusters in a breast cancer study, including in situ and invasive cancer subgroups.
• Distinguished adipose from connective tissues, a capability lacking in other methods.

Conclusions:

• TransST is an effective and robust method for spatial transcriptomics data analysis.
• The framework accurately identifies cell subclusters and their associated biomarkers.
• TransST advances the capability to interpret complex spatial transcriptomic data.