Xuegong Zhang

Forecasting cell state futures from static snapshots

This paper introduces CellTempo, a framework that forecasts long-range cell state trajectories from static transcriptomic data, demonstrating its ability to predict lineage progression and distinguish perturbation responses from long-term fate changes.

OmicOS: A Comprehensive Omics Ecosystem Infrastructure and Agent System for the AI Era

OmicOS provides an agent-operable omics ecosystem that enables AI systems to execute and verify analyses using community methods, achieving high performance in benchmark tasks.

Multi-modal Diffusion Model with Dual-Cross-Attention for Multi-Omics Data Generation and Translation

scDiffusion-X generates high-fidelity multi-omics data and enables modality translation with uncertainty quantification and gene regulatory network inference.

How different AI models understand cells differently

This paper introduces scGeneLens, a framework to analyze how different AI models perceive cells, revealing that scFoundation focuses on cell-type markers while scGPT emphasizes shared pathways.

ERNIE-RNA: An RNA Language Model with Structure-enhanced Representations

ERNIE-RNA is an RNA language model that incorporates base-pairing restrictions to improve structural representation, achieving state-of-the-art performance in RNA structural and functional predictions.

Unified multimodal learning enables generalized cellular response prediction to diverse perturbations

X-Pert is a unified framework for predicting cellular responses to diverse perturbations, outperforming existing methods across genetic and chemical perturbations and enabling generalization across cell types.

Tokenizing single-cell transcriptomes as a native language for large language models

This paper introduces CellTok, a method to convert single-cell transcriptomes into tokens for LLMs, enabling joint processing of cellular data and text.

SpatialZoomer: multi-scale feature analysis of spatial transcriptomics

SpatialZoomer identifies disease progression signals and spatially heterogeneous niches in spatial transcriptomics data by analyzing features at multiple scales.

GeST: Towards Building A Generative Pretrained Transformer for Learning Cellular Spatial Context

GeST is a generative transformer model pretrained on spatial transcriptomics data to learn spatial context, showing improved generation and zero-shot spatial niche embedding extraction.

scRecover: Discriminating true and false zeros in single-cell RNA-seq data for imputation

scRecover imputes dropout zeros in scRNA-seq data while preserving true zeros, improving downstream analysis through accurate zero discrimination.

uniLIVER: a Human Liver Cell Atlas for Data-Driven Cellular State Mapping

uniLIVER creates a human liver cell atlas using sc/snRNA-seq data and develops LiverCT to map abnormal datasets to the reference atlas, revealing changes in cell type composition and states in liver cancer.

Benchmarking AI Models for In Silico Gene Perturbation of Cells

This study benchmarks 10 in silico gene perturbation methods across four scenarios, identifying strengths and limitations in predicting perturbation effects and zero-shot transfer.

CellMemory: Hierarchical Interpretation of Out-of-Distribution Cells Using Bottlenecked Transformer

CellMemory interprets out-of-distribution cells using a bottlenecked Transformer, outperforming pre-trained models and enabling accurate analysis of spatially resolved single-cell data.

A universal single-cell transcriptomics atlas of human lung decodes multiple pulmonary diseases

The paper presents uniLUNG, a large single-cell transcriptomic atlas of the human lung, identifying new cell types and subpopulations associated with lung diseases.

Building a learnable universal coordinate system for single-cell atlas with a joint-VAE model

The paper introduces UniCoord, a joint-VAE model that learns a universal coordinate system for single-cell transcriptomic data to capture heterogeneity and enable data generation and analysis.

Large Scale Foundation Model on Single-cell Transcriptomics

This paper introduces scFoundation, a large-scale pretrained model for single-cell transcriptomics, achieving state-of-the-art results in multiple downstream tasks.

Toward mastering the cell language by learning to generate

The paper introduces scMulan, a large AI model trained to generate cell transcriptomics data with specific attributes, showing it can generate pseudo data, predict missing attributes, and identify driving regulators.

Benchmarking multi-omics integration algorithms across single-cell RNA and ATAC data

This study benchmarks 12 multi-omics integration methods on scRNA-seq and scATAC-seq data across three tasks, identifying method-specific strengths and providing selection guidelines.

scDecouple: Decoupling cellular response from infect-ed proportion bias in scCRISPR-seq

scDecouple separates true cellular response from gRNA infection bias in scCRISPR-seq data, improving gene perturbation analysis in heterogeneous samples.

scDirect: key transcription factor identification for directing cell state transitions based on single-cell multi-omics data

scDirect identifies key transcription factors driving cell state transitions using single-cell multi-omics data, showing robust performance in benchmarking and real-world applications.

Hidden Knowledge Recovery from GAN-generated Single-cell RNA-seq Data

The paper generates synthetic single-cell RNA-seq data using a custom GAN and recovers hidden developmental knowledge, outperforming existing models on limited datasets.

PAST: latent feature extraction with a Prior-based self-Attention framework for Spatial Transcriptomics

PAST characterizes spatial domains in spatial transcriptomics data using a variational graph convolutional auto-encoder with prior-based self-attention, improving downstream analyses.

Decoding functional cell-cell communication events by multi-view graph learning on spatial transcriptomics

This paper introduces HoloNet, a method to decode functional cell-cell communication events from spatial transcriptomic data by integrating ligand-receptor pairs, cell-type spatial distribution, and downstream gene expression.

Differentiation and migration of hematopoietic stem and progenitor cells cross multiple tissues

The paper integrates single-cell RNA-seq data from multiple tissues to construct a cell atlas of hematopoietic stem and progenitor cells (HSPCs) and identifies migration patterns and signaling interactions involved in their cross-tissue differentiation.

simCAS: an embedding-based method for simulating single-cell chromatin accessibility sequencing data

simCAS generates high-fidelity scCAS data using embeddings, enabling benchmarking of computational methods for cell clustering, trajectory inference, and cis-regulatory interaction analysis.

A generic reference defined by consensus peaks for scATAC-seq data analysis

The paper constructs a generic chromatin accessibility reference (cPeaks) from bulk ATAC-seq data to improve scATAC-seq analysis and cell type identification.

xTrimoGene: An Efficient and Scalable Representation Learner for Single-Cell RNA-Seq Data

xTrimoGene is a scalable transformer model for scRNA-seq data that reduces computational demands while maintaining accuracy, achieving state-of-the-art results in downstream tasks.

STEM: A Method for Mapping Single-cell and Spatial Transcriptomics Data with Transfer Learning

STEM maps single-cell and spatial transcriptomics data using transfer learning to create spatially aware embeddings and predict pseudo-spatial adjacencies.

hECA: the cell-centric assembly of a cell atlas

The paper introduces hECA, a cell-centric human Ensemble Cell Atlas, which integrates scRNA-seq data from multiple studies to enable targeted data retrieval, multi-view presentations, and customizable reference creation for cell type annotations.

Discovering single-cell eQTLs from scRNA-seq data only

The paper introduces eQTLsingle, a method to discover single-cell eQTLs using only scRNA-seq data, revealing cell-type-specific tumor-related eQTLs not found in bulk studies.