Fusion transcript detection using spatial transcriptomics

Area of Science:

• Oncology
• Molecular Biology
• Genomics

Background:

• Fusion transcripts are key drivers of tumor heterogeneity, evolution, and treatment resistance.
• Studying fusion transcripts at high spatial resolution in tissues has been limited by the lack of full-length transcript data.
• Spatial transcriptomics offers near single-cell resolution but doesn't directly detect fusion transcripts.

Purpose of the Study:

• To develop a novel method for inferring and spatially localizing fusion transcripts in tissue sections.
• To leverage spatial transcriptomics data to overcome limitations in fusion transcript detection.
• To enable high-resolution spatial analysis of fusion transcripts in cancer.

Main Methods:

• Introduced STfusion, a method utilizing spatial transcriptomics to infer poly(A) tail presence/absence.
• Developed a C-score to quantify differences between observed and expected poly(A) tail counts.
• Applied STfusion to HeLa cells with known fusions and clinical prostate cancer data.

Main Results:

• STfusion successfully predicted fusion transcripts in cell lines and clinical samples.
• The spatial distribution of the SLC45A3-ELK4 cis-SAGe was mapped in prostate cancer tissues at near single-cell resolution.
• The cis-SAGe was found in cancerous, pre-neoplastic, and inflamed regions, as well as normal glands.

Conclusions:

• STfusion effectively detects and spatially localizes fusion transcripts in cancer.
• The method distinguishes true fusion transcripts from trans-splicing events.
• Enables high-resolution spatial analysis of fusion transcripts in clinical tissue sections.