Clonal phylogenies inferred from bulk, single cell, and spatial transcriptomic analysis of epithelial cancers

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Epithelial cancers, like prostate cancer, exhibit significant histological and genomic heterogeneity.
• Tumor genetics studies reveal extensive inter- and intra-patient genomic variation.
• Machine learning advances allow inferring genomic single-nucleotide variant (SNV) and copy number variant (CNV) status from transcriptomic data.

Purpose of the Study:

• To assess the accuracy of transcript-based inferred tumor phylogenies in recapitulating DNA-based phylogenies.
• To compare inferred SNV and CNV phylogenies with ground-truth DNA phylogenies.
• To evaluate the utility of transcriptomics for reconstructing cancer evolutionary history.

Main Methods:

• In-silico comparison of inferred and directly resolved SNV and CNV status from single cancer cells across three cell lines.
• Analysis of published prostate cancer DNA phylogenies against inferred transcript-based phylogenies.
• Comparison of pseudo-bulked spatial transcriptomic data with whole-genome sequencing (WGS) data from adjacent tissue sections.

Main Results:

• Inferred SNV phylogenies accurately recapitulated DNA phylogenies with low entanglement (0.097).
• Inferred copy number variant (iCNV) and CNV-based phylogenies also showed high accuracy (entanglement = 0.11).
• Phylogenetic concordance was observed between published DNA phylogenies and inferred transcript-based phylogenies, including spatial transcriptomics data (entanglement = 0.35).

Conclusions:

• Transcript-based inferred phylogenies effectively recapitulate conventional DNA-based tumor phylogenies.
• This approach offers a viable method for studying cancer evolution using transcriptomic data.
• Future research should focus on enhancing the accuracy, genomic, and spatial resolution of these transcript-based phylogenies.