KLONE: An Imbalance-Robust k-NN Ensemble for Cross-Modal Single-Cell Clone Assignment

KLONE assigns unpaired scRNA-seq cells to scDNA-seq copy-number clones with highest accuracy on three colorectal cancer patients, achieving 100% recall of minority clones.

• Why it matters: Cross-modal clone assignment is hard because scDNA-seq and scRNA-seq profile different cells and only a weak, gene-dependent dosage signal links modalities, with no RNA clone labels.
• What they did: KLONE uses a per-clone k-nearest-neighbor ensemble based on cross-modal Pearson similarity and gives each clone an identical neighborhood budget to avoid majority-clone dominance.
• The result: Compared with six existing methods, KLONE most accurately matches RNA clone labels and recovers the RNA clonal distribution almost perfectly with ρ = 0.99.

AI summary of the abstract · may contain errors · not medical advice · read the paper.