About Me
I am Wenduo Cheng, a 3rd-year Ph.D. student in the Ray and Stephanie Lane Computational Biology Department at Carnegie Mellon University. I am fortunate to be advised by Jian Ma. Before my Ph.D., I completed an M.S. in Computational Biology at CMU and a B.S. in Genetics and Genomics at Duke Kunshan University.
My background spans bioinformatic sequence analysis, evolutionary biology, environmental science, genome-wide association studies, and deep learning for computational biology. My current research interests lie in applying AI to accelerate biological discovery, particularly through biological foundation models and autonomous agents. My work centers on these directions:
- Biological Foundation Models: I build biologically meaningful โvirtual cellโ models that capture evolutionary and regulatory constraints, aiming for representations that generalize across cell types, modalities, and biological contexts.
- Autonomous Agents for Science: I develop LLM agents that accelerate experimental design, computational analysis, and lab automation, turning open-ended scientific questions into reproducible, executable workflows.
- Gene Regulation & Multimodal Modeling: I study the principles of gene regulation through multimodal modeling, integrating sequence, structure, and functional signals to reveal how regulatory information is encoded and read out.
๐ฅ News
08/2026: ๐ Our perspective Towards Human-Led, Agent-Driven Autonomous Laboratories for the Life Sciences is now available!
06/2026: ๐ Our SKILLFOUNDRY paper is accepted by COLM 2026!
05/2026: ๐ผ I will be joining Genentech as an intern starting 05/2026!
05/2026: ๐ Our AgentCo-op preprint is now available on arXiv!
11/2025: ๐ DNALongBench has been featured for the Nature Communications Editorsโ Highlights collection โ Computational and Theoretical Biology!
09/2025: ๐ Our DNALongBench work is accepted by Nature Communications!
08/2025: ๐ Our L2G paper is accepted by TMLR 2025!
๐ Selected Publications
A full list of publications is available here. (โ indicates equal contribution.)
-
Towards Human-Led, Agent-Driven Autonomous Laboratories for the Life SciencesPreprints.org, 2026. -
AgentCo-op: Retrieval-Based Synthesis of Interoperable Multi-Agent WorkflowsarXiv preprint arXiv:2605.20425, 2026. -
SKILLFOUNDRY: Building Self-Evolving Agent Skill Libraries from Heterogeneous Scientific ResourcesConference on Language Modeling (COLM), 2026. -
Cyanobacterial Blooms Are Not a Result of Positive Selection by Freshwater EutrophicationMicrobiology Spectrum 12(3), e03194โ22, 2022.

