Yuying Xie

InsilicoCell predicts cellular functional profiles and enables multi-objective virtual drug screening by integrating diverse molecular data and tasks.

GEOMeta automates metadata curation for GEO transcriptomic data, generating standardized annotations for 600,000 human bulk RNA-seq samples and benchmarking LLMs for annotation quality.

The paper presents a longitudinal multimodal benchmarking dataset and competition to evaluate modality prediction methods for gene regulatory inference, showing top models outperform existing approaches.

scLinguist translates between single-cell multi-omics modalities by pretraining on unimodal data and fine-tuning on paired data, outperforming existing methods in predicting protein abundance from RNA.

TESSA detects trajectory-preserved and spatially-variable genes in spatial transcriptomics, improving tumor progression analysis and spatial domain detection.

DANCE 2.0 automates and interprets single-cell preprocessing by recommending optimal pipelines for various tasks, improving performance across six major analyses.

This study identifies IL-1β-inducing pathobionts in Crohn's disease that drive inflammation and may predict response to anti-TNF therapies.

This study identifies a SOX2-CCL2-IL1 pathway that drives immune suppression in precancerous epithelial dysplasia, promoting malignant transformation.

This paper introduces a new method for clustering and visualizing single-cell RNA-seq data using path metrics, which outperforms existing methods in preserving both global geometry and cluster structure.

DANCE provides a benchmark platform for single-cell analysis with 32 methods across 21 datasets, enabling easy reproduction and development of deep learning models.

The paper introduces a GAN-based method to generate high-quality membranes for cells with poor staining in fluorescent microscopy images, improving segmentation accuracy.

This paper introduces a general single-cell analysis framework using conditional diffusion models, showing competitive performance across tasks with added flexibility via prior information.

Treating cells as tokens and tissues as sentences, with spatial transcriptomics added during pre-training, gives a single-cell foundation model the authors report outperforms existing models and runs far faster.

The paper introduces a large-scale spatial transcriptomic dataset and a Graph Neural Network-based method for cell type deconvolution in human immuno-oncology.

supCPM improves scRNA-seq visualization by preserving global structure and cluster variance using supervised clustering guidance.

FARDEEP improves tumor-infiltrating lymphocyte deconvolution by using least trimmed squares to reduce outliers, providing more accurate immune cell composition estimates.