Kun Huang

ARSA: an autonomous research scientist for target nomination in Alzheimer's disease and related dementias

ARSA generates evidence-based target shortlists for Alzheimer's disease and related dementias by integrating natural language research interests with molecular data and cross-cohort evidence.

Patient-derived tau-seeded human neuronal chimeras recapitulate mature Alzheimer's tau pathology and uncover human-specific neuronal vulnerability

This study develops a human neuronal chimera model to study Alzheimer's tau pathology, showing human-specific vulnerability and recapitulating mature AD features in vivo.

From time-course expression to gene regulation: direct linear ODE inference without finite-difference approximation

Fitting a linear ODE with an equilibrium term via its matrix-exponential solution, not finite differences, raises TF-network AUROC over competing methods on simulated and scRNA-seq data.

Identification of High-Risk Cells in Single-Cell Spatially Resolved Transcriptomics Data Using DEGAS Spatial Smoothing

The paper introduces DEGAS, a method to identify high-risk cells in spatial transcriptomics data by transferring disease attributes to individual cells, validated in cancer and diabetes datasets.

Building explainable graph neural network by sparse learning for the drug-protein binding prediction

This paper introduces SLGNN, a sparse learning approach for drug-protein binding prediction that identifies chemically valid key structures with better predictive power than existing methods.

A novel deep optimal transport framework reveals prostate cancer risk heterogeneity, Alzheimers disease risk heterogeneity, and myeloma cells associated with a short-term sham bortezomib response and progressive disease

The paper introduces DEGASv2, a deep learning framework for analyzing single-cell and spatial transcriptomics data, revealing disease risk heterogeneity and drug response patterns in multiple conditions.

Identify Alzheimer's disease subtypes and makers from multi-omic data of human brain and blood with a subspace merging algorithm

This study integrates multi-omic data from brain and blood to identify AD subtypes and proposes LDLR as a blood biomarker linked to brain gene expression changes.

Diagnostic Evidence Gauge of Spatial Transcriptomics (DEGAS): Using transfer learning to map clinical data to spatial transcriptomics in prostate cancer

The paper introduces DEGAS, a transfer learning framework that maps clinical data to spatial transcriptomics in prostate cancer, identifying transcriptional signatures of carcinogenesis in morphologically normal tissues.

AIscEA: Unsupervised Integration of Single-cell Gene Expression and Chromatin Accessibility via Their Biological Consistency

AIscEA integrates single-cell gene expression and chromatin accessibility data by aligning cell types based on biological consistency, showing improved robustness and accuracy over existing methods.

SPCS: A Spatial and Pattern Combined Smoothing Method of Spatial Transcriptomic Expression

This paper introduces SPCS, a two-factor smoothing method for spatial transcriptomics data that improves separability and biological interpretability compared to one-factor methods.

Diagnostic Evidence GAuge of Single cells (DEGAS): A flexible deep-transfer learning framework for prioritizing cells in relation to disease

DEGAS transfers disease information from patients to cells using deep transfer learning, identifying PHF19high myeloma cells linked to progression.

A deep learning approach for tissue spatial quantification and genomic correlations of histopathological images

The paper develops a CNN model to classify epithelial and stromal tissues in histopathological images and correlates these with gene expression data to identify biological processes.

MORONET: Multi-omics Integration via Graph Convolutional Networks for Biomedical Data Classification

MORONET is a multi-omics integration method using graph convolutional networks for biomedical classification, outperforming existing approaches and identifying disease-related biomarkers.

WEVar: a novel statistical learning framework for predicting noncoding regulatory variants

WEVar integrates functional scores from multiple methods to predict noncoding regulatory variants, outperforming existing approaches in simulation and empirical studies.

TSUNAMI: Translational Bioinformatics Tool Suite For Network Analysis And Mining

TSUNAMI is a tool for mining gene co-expression networks and performing downstream analysis, offering features like web interface, database access, and visualization.

Integration of Mouse and Human Single-cell RNA Sequencing Infers Spatial Cell-type Composition in Human Brains

The paper integrates mouse and human single-cell RNA sequencing data to predict and spatially map cell-type composition in human brain tissue.

Attenuation of epigenetic regulator SMARCA4 and ERK-ETS signaling suppresses aging-related dopaminergic degeneration

The paper identifies SMARCA4 and ERK-ETS signaling as targets for suppressing aging-related dopaminergic degeneration in Parkinson's disease using gene co-expression analysis and Drosophila models.

A semi-supervised approach for cell phenotypic and functional estimation in tissue microenvironment

ICTD improves tissue microenvironment analysis by enhancing cell type identification, marker gene discovery, and functional inference through a semi-supervised deconvolution approach.

BERMUDA: A novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes

BERMUDA corrects batch effects in scRNA-seq data and reveals hidden cellular subtypes by transferring information across batches.

LAmbDA: Label Ambiguous Domain Adaption Dataset Integration Reduces Batch Effects and Improves Subtype Detection

LAmbDA reduces batch effects and improves subtype detection by integrating datasets with ambiguous labels across species and studies.