James Zou

The paper introduces a metric to measure surface protein polarity from immunofluorescence imaging and shows it identifies immune cell states linked to tumor engagement, improving survival prediction models.

The paper introduces MIISS, a framework for similarity search in multiplexed immunofluorescence images, showing domain-specific models outperform others.

GenePT embeds National Center for Biotechnology Information (NCBI) gene text descriptions using GPT-3.5 and aggregates them by expression to represent single cells without pretraining on millions of transcriptomes.

This work introduces SCGP, a method for unsupervised identification of tissue structures from spatial omics data, and demonstrates its application to spatial transcriptomics and proteomics datasets.

This paper introduces SMAI, a framework for alignability testing and integration of single-cell data, showing it outperforms existing methods and improves downstream analyses.

7-UP generates in silico 40-plex CODEX data from a 7-plex mIF panel using cellular morphology, enabling cell type classification and survival prediction across diverse samples.

The paper builds a visual-language foundation model for pathology AI using a large dataset of pathology images and natural language descriptions curated from medical Twitter.

The paper introduces a spectral method to evaluate and combine multiple data visualizations, improving the quality of the consensus visualization by leveraging eigenscores.

This study examines the impact of random external factors on eLife peer review outcomes, finding that such factors have relatively little quantifiable bias.

DynamicViz generates dynamic visualizations to assess the reliability of dimensionality reduction methods by quantifying dynamic variability in data.

The paper predicts immune escape by using pretrained protein language model embeddings, showing they outperform zero-shot methods but are not superior to simple baselines.

SPACE-GM identifies spatial motifs in cancer samples linked to recurrence and immunotherapy outcomes using geometric deep learning.

The paper identifies proteomic features of resilience to Alzheimer's disease, finding lower soluble Aβ levels and specific protein interactions in brain regions associated with resistance to dementia.

This paper introduces scDRS, a method to identify disease-associated cells in scRNA-seq data using polygenic risk scores, finding novel subpopulations linked to diseases like IBD and schizophrenia.

The paper introduces computational tools to analyze crosslink-ligation RNA data, revealing new RNA structures and homodimers.

DynaMorph uses self-supervised learning to analyze morphodynamic states of live human microglia, identifying distinct states and transitions in response to disease-relevant stimuli.

This paper introduces a method to combine histology and spatial gene expression data to infer high-resolution transcriptome maps without sequencing or hybridization.

TCR-BERT uses self-supervised transfer learning on unlabeled T-cell receptor sequences to learn representations that improve antigen-binding prediction, specificity clustering, and computational TCR sequence design.

BABEL translates between scRNA-seq and scATAC-seq at single-cell resolution, enabling synthesis of paired multi-omic data when only one modality is available.

The paper predicts SARS-CoV-2 RNA localization to host mitochondria and nucleolus using machine learning, identifying key regions in viral RNA that drive this localization.

The paper highlights the lack of standard definitions and protocols for collecting and using race, ethnicity, and ancestry data in clinical genetics and precision medicine.

The paper analyzes the Tabula Muris Senis dataset to identify aging-related gene expression changes across 76 tissue-cell types in mice, revealing both shared and cell-type specific aging signatures.

The paper develops an automated method for RNA secondary structure analysis using network techniques, validated against manual analysis and known structures.

The paper develops a tool to analyze spatial gene expression correlations in tissue sections, applied to human fetal pancreas development.

AdaFDR improves multiple hypothesis testing by using covariates to increase discovery power while controlling false discovery rate, outperforming BH and IHW in eQTL analysis.

The paper analyzes REA data collection practices and disparities in gnomAD, highlighting the need for standardized frameworks to improve diversity and variant interpretation in clinical genomics.

This study systematically characterizes genome editing in primary T cells, identifies proximal genomic insertions, and develops a machine learning model to predict CRISPR-Cas9 DNA repair outcomes.