Tae Hyun Hwang

ORBIS reconstructs 3D tissue geometry from single-cell gene expression without spatial references, achieving lower shape error and faster sampling than existing methods.

This study shows that ARID1A loss in gastric cancer affects immune evasion depending on tumor context, with different outcomes in mouse models and human subtypes.

UniST reconstructs dense 3D spatial transcriptomics data from sparse serial sections, restoring structural continuity and biologically meaningful expression patterns in multiple tissue contexts.

UTOPIA quantifies confidence in virtual spatial transcriptomics predictions across scales, improving reliability and preventing false conclusions.

Pixel2Gene improves spatial gene expression reconstruction and prediction by integrating histology images with spatial transcriptomics data.

REMAP reconstructs spatial tissue architecture from scRNA-seq using deep learning and ST references, revealing microenvironmental features and cancer subtypes.

HistoSweep improves tissue quality control in gigapixel images by removing artifacts and preserving biologically relevant structures, enhancing analysis accuracy in digital pathology and spatial omics.

S2Omics automatically selects regions of interest from histology images to maximize molecular information content in spatial omics experiments, improving reproducibility and downstream biological discovery.

iSCALE predicts super-resolution gene expression and annotates tissue architecture in large tissues, revealing features missed by conventional spatial transcriptomics.

This paper develops a 3D virtual H&E staining method using holotomography and deep learning for cancer tissues, validated on colon and gastric cancer samples.

This study shows that TMPRSS11B promotes an acidified and immunosuppressive tumor microenvironment in lung squamous cell carcinoma, suggesting it as a therapeutic target.

MSI-SEER uses deep Gaussian processes to predict microsatellite instability from histology images, achieving high accuracy in MSI prediction and ICI response prediction.

The paper uses whole slide images to analyze spatial heterogeneity of tumor mutation burden and immune infiltrates in bladder cancer, finding that low spatial heterogeneity with high TILs correlates with improved survival.

This paper uses deep learning to analyze histopathology images of colorectal cancer to predict progression-free survival by quantifying tumor infiltrating lymphocytes.

This paper introduces a Bayesian semi-nonnegative matrix tri-factorization method to identify cancer-associated pathways from real-valued gene expression data, showing improved performance over existing methods.

This study shows that CRISPR/Cas9-induced INDELs can lead to mRNA misregulation by creating foreign mRNAs or proteins in about half of edited cell lines.

This paper develops a CNN-based method to analyze TILs in bladder cancer WSIs, showing their fraction and spatial organization correlate with patient survival.

This paper introduces a texture descriptor called SLBP for prostate cancer Gleason score prediction from whole slide images, achieving over 80% accuracy in distinguishing grades ≤7 vs ≥8.

This paper develops a Gaussian process-based method for TMB prediction from whole slide images, outperforming baseline methods by incorporating heteroscedastic noise modeling.

This paper presents an automatic method for Gleason score classification of prostate cancer using CSLBP descriptors and SVM, achieving over 79% accuracy on TCGA data.

The paper develops a 39-gene signature to stratify anaplastic gliomas into two subgroups with distinct survival outcomes, validated in an external dataset.

This study shows that miR-16 overexpression suppresses growth of rhabdoid tumor cell lines by reducing proliferation and increasing apoptosis.

This study identifies cancer-type-specific altered pathways linked to survival outcomes across 19 cancer types using integrated somatic mutation and gene expression data.