Ziwei Yang, Ph.D.

Hi, I’m Ziwei Yang. I am a Researcher at the Bioinformatics Center, Institute for Chemical Research, Kyoto University, and a Bioinformatics Scientist at Keiji AI.

I work at the intersection of computational biology and artificial intelligence, developing AI systems for understanding, modeling, and engineering biological systems.

My research connects two directions:

  • Mechanistic Biological Modeling — combining genome-scale models, biological knowledge, and data-driven learning to model and engineer complex biological systems.
  • AI Agents for Scientific Discovery — building intelligent agents that integrate multimodal biological data, reason over scientific knowledge, and support biomedical research.

Currently bridging data-driven AI with mechanistic biological modeling for metabolic engineering and biomedical discovery.

  1. HackathonBarcelona

    Participating in BioHackathon Europe 2026

  2. ConferenceBali

    Presenting CrossGapFilling at iscbAI 2026

  3. Publication

    Published DeepEvidence in Nature Machine Intelligence

Research Focus

All projects →
  1. 01Model Biology

    Mechanistic Biological Modeling

    Developing data-driven and graph-based methods for genome-scale metabolic models, with a focus on growth-coupled gene deletion and metabolic engineering.

    GraphGDelDeepGDelDBgDelPhysics-Constrained Models
  2. 02Understand Biology

    Multimodal Biomedical AI

    Learning representations from multi-omics and other biomedical data to characterize biological systems and disease heterogeneity.

    MLOmicsGeSubNetMultimodal Biological Intelligence
  3. 03Discover Biology

    AI Agents for Scientific Discovery

    Developing AI agents that can reason over biological knowledge, integrate heterogeneous biomedical data, and support reproducible scientific discovery.

    DeepEvidenceBiomedical AI CopilotsAutonomous Biomedical Research

Selected Work

All publications →

Making Large Language Models Reliable Data Science Programming Copilots for Biomedical Research

Zifeng Wang, Benjamin Danek, Ziwei Yang, Zheng Chen, Jimeng Sun

Nature Biomedical Engineering

Benchmarking 16 LLMs on 293 coding tasks from 39 biomedical studies shows overall accuracy below 40%; an agent that first drafts and iteratively refines an analysis plan reaches 74%.

Empowering Biomedical Evidence Exploration and Synthesis with Deep Knowledge Graph Research

Zifeng Wang, Zheng Chen, Ziwei Yang, Xuan Wang, Qiao Jin, Yifan Peng, Zhiyong Lu, Jimeng Sun

Nature Machine Intelligence

DeepEvidence, a multi-agent deep research system that searches biomedical knowledge bases and literature, builds an evidence graph as it goes, and outperforms generic AI agents on four open benchmarks.

GraphGDel: Constructing and Learning Graph Representations of Genome-Scale Metabolic Models for Growth-Coupled Gene Deletion Prediction

Ziwei Yang, Takeyuki Tamura

IEEE Transactions on Computational Biology and Bioinformatics

Builds graph representations of genome-scale metabolic models and combines them with gene and metabolite sequences to predict growth-coupled gene deletion strategies, outperforming baselines on three metabolic models.

MLOmics: Cancer Multi-Omics Database for Machine Learning

Ziwei Yang, Rikuto Kotoge, Xihao Piao, Zheng Chen, Lingwei Zhu, Peng Gao, Yasuko Matsubara, Yasushi Sakurai, Jimeng Sun

Scientific Data

An open cancer multi-omics database for machine learning: 8,314 patients across 32 cancer types with four omics types, stratified feature versions and baseline benchmarks.

Oral

GeSubNet: Gene Interaction Inference for Disease Subtype Network Generation

Ziwei Yang, Zheng Chen, Xin Liu, Rikuto Kotoge, Peng Chen, Yasuko Matsubara, Yasushi Sakurai, Jimeng Sun

International Conference on Learning Representations (ICLR)

Joins a deep generative model of patient expression profiles with a graph neural network over prior gene networks to generate subtype-specific gene interaction networks, with average gains of 20–57% on four graph metrics across four cancer datasets.