scVAR: integrating genomics and transcriptomics from single-cell RNA-seq -insights from leukemia case studies

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• High-throughput technologies, including single-cell RNA sequencing (scRNA-seq), are revolutionizing biomedical research by enabling detailed analysis of cellular heterogeneity.
• Complex diseases like Acute Myeloid Leukemia (AML) and Acute Lymphoblastic Leukemia (ALL) exhibit significant genetic and phenotypic diversity, complicating diagnosis and treatment.
• While DNA sequencing is standard for genetic variation, the transcriptome also holds valuable genomic information.

Purpose of the Study:

• To introduce scVAR, a novel computational framework designed to learn and integrate genetic variation directly from scRNA-seq data.
• To enhance the detection of subtle cellular differences by unifying transcriptomic and variant-derived information using a cross-attention mechanism.
• To demonstrate the utility of scVAR in leukemia case studies for improved cell identity and subpopulation discovery.

Main Methods:

• Development of scVAR, a computational framework utilizing variational autoencoders.
• Implementation of a paired encoder-decoder architecture with a cross-attention-based fusion layer.
• Application of scVAR to analyze scRNA-seq data from leukemia case studies.

Main Results:

• scVAR successfully integrates transcriptomic and variant data, revealing cell identities missed by analyzing each data type separately.
• The framework identified 20%-30% more subpopulations compared to transcriptomic analysis alone, even with limited variant coverage.
• Variant detection was optimized within captured regions for 3' scRNA-seq, maximizing information extraction.

Conclusions:

• scVAR effectively bridges the gap between transcriptomics and genomics for single-cell analysis.
• The framework provides a broadly applicable platform for the integrative characterization of cell states and disease processes.
• Integrating variant information with scRNA-seq data significantly enhances the resolution of cellular heterogeneity in complex diseases.