Fan Zhang

OTTER improves bisulfite deconvolution in tumor xenograft sequencing by restoring recall to 52%-54% while preserving transcriptomic accuracy.

This study identifies an AMPK-dependent brainstem circuit that promotes recovery from metabolic coma in mice through widespread norepinephrine release.

CellDot improves spatial transcriptomics by accurately decontaminating transcript assignments using optimal transport, enhancing biological interpretation and spatial pattern recovery.

Oakuloid combines a structural machine learning model with a 3D liver-on-chip assay to predict drug-induced liver injury, achieving high AUROC on external benchmarks and offering orthogonal confirmation signals.

This study identifies CAD as a key regulator of myocardial ischemia/reperfusion injury by linking nucleotide metabolism to oxidative stress through dual-compartment ROS signaling.

STAT is a multi-agent framework that improves spatial transcriptomics analysis by enabling interactive, transparent, and efficient multi-task analysis across different data resolutions.

This paper expands the genetic code by incorporating non-canonical monomers into proteins using evolved tRNAs and orthogonal synthetases, enabling broader sequence context compatibility.

DrugPlayGround benchmarks large language models for drug discovery tasks like drug characteristics and interactions, assessing their performance against traditional methods.

The paper develops a thermostable green fluorescent protein, Matcha, for live imaging in hyperthermophiles and uses it to study division ring dynamics in Sulfolobus.

This study uses deep learning and genome-wide association studies to identify pest-resistant genes in grapevine, achieving high accuracy in predicting pest damage.

DBSCAN-SWA rapidly detects and annotates prophages in bacterial genomes with high recall and user-friendly visualizations.