Rui Kuang

CTFacTomo reconstructs 3D spatial gene expressions from RNA tomography data using tensor factorization and regularization, outperforming benchmarks on zebrafish and mouse datasets.

This paper predicts spatial gene expression from histological images using adaptive spatial GNNs, improving performance on breast cancer data.

BiGCN improves single-cell RNA-seq imputation by integrating cell similarities and gene co-expression using dual graph convolutional networks, outperforming existing methods in imputation and clustering.

FIST-nD imputes 2D and 3D spatial transcriptomics data using graph-regularized tensor completion, improving accuracy for downstream analyses.

This paper introduces FIST, a graph-regularized tensor completion model for imputing missing mRNA expressions in spatial transcriptomics data, showing improved performance over existing methods.

The paper introduces KPI, a method to accurately infer KIR genes and haplotypes from genome sequencing reads using novel k-mer signatures, achieving high accuracy on synthetic and real data.

The paper generates and aligns KIR haplotypes at high resolution, defines them as 14 loci with 9 genes, and applies the alignments in software tools for haplotype annotation and genotyping.