ONE TREE, MANY CELLS: SINGLE-CELL DATA INTEGRATION ACROSS SPECIES
Track: Tiny Paper Track
Keywords: scRNA-Seq, data integration, joint embeddings, Cross-species, Astrocytes
TL;DR: There is a critical need for novel computational models for cross-species single cell data integration.
Abstract: Meaningfulness Statement: In this work we highlight a fundamental question in comparative genomics: How can brain cell populations, characterized at single-cell resolution, be meaningfully compared across distantly related species? While scRNA-Seq now enables the molecular characterization of complex tissues across diverse species, existing integration methods fail to capture biologically meaningful patterns across taxa due to critical differences in gene orthology. We assessed current integration models in aligning brain cell populations across human, macaque, mouse, fruit fly, and honey bee, highlighting the need for improved models. We propose using orthogroups—sets of genes from a common ancestor—and transcriptional modules to enhance integration.
Submission Number: 96
Loading