Jinman Kim

GHIST+ reconstructs tissue-wide molecular maps from H&E histology, enabling cohort-scale molecular reconstruction from sparse spatial transcriptomic data.

This study evaluates how large language models generate summaries of bioinformatics analyses, finding they produce readable and safe summaries but often misinterpret visualizations and lack novel insights.

GHIST predicts spatial gene expression at single-cell resolution from histology images, showing superior performance across different resolutions and datasets.

This study benchmarks six methods for predicting spatial gene expression from histology images using breast and skin cancer datasets and TCGA data, evaluating their performance and translational potential.

BIDCell improves subcellular spatial transcriptomics segmentation by combining self-supervised learning with biologically informed loss functions, outperforming existing methods across multiple tissue types.

This paper introduces a deep multimodal graph-based network to predict breast cancer survival by integrating spatial phenotype data from imaging mass cytometry with patient variables.