Anrui Wang

Anrui Wang

CS undergraduate · AI for Science

[ShanghaiTech University]

I'm an undergraduate computer science student working at the intersection of machine learning and biology — AI for Science. I'm drawn to problems where learning the right representation of noisy, high-dimensional biological data turns into predictions you can actually trust. My current projects live in single-cell genomics and molecular networks, but my interests run broadly across machine learning for the life sciences, and I'm looking for research opportunities and collaborations.

  • Jan 2026Began research with Prof. Xuming He on topology-constrained deep learning for inductive interactome reconstruction.
  • Nov 2025Started research with Prof. Jie Zheng on multi-omics modeling for synthetic-lethality prediction.

Synthetic-Lethality Prediction via STATE-Based Multi-Omics Modeling

Supervised by Prof. Jie Zheng · Nov 2025 – present

Predicting synthetic-lethal gene pairs for cancer-selective drug-target prioritization by integrating Perturb-seq perturbation transcriptomics, DepMap functional dependency (CRISPR GeneEffect), and ESM2 protein-sequence embeddings. Models perturbation responses as cell-state representations via a prototype-based distribution-regression approach that recovers dependency signal missed by pseudobulk and attention baselines.

Perturb-seqDepMapscVIESM2distribution regression

Topology-Constrained Inductive Interactome Reconstruction

Supervised by Prof. Xuming He · Jan 2026 – present

Reframes protein–protein interaction prediction from independent pairwise classification into inductive reconstruction of the full interaction network, using a topology-constrained conditional generator (pairwise compatibility, hub propensity, community membership, set-level density) distilled from a masked-edge topology teacher while staying strictly inductive on intrinsic ESM protein features.

GNNgraph generationtopologyESM
Programming
Python, PyTorch
ML / Tooling
HuggingFace Accelerate, single-cell foundation models, ESM series, graph neural networks
Coding Agents
Claude Code, Codex
Domains
single-cell genomics, cancer dependency (DepMap), synthetic lethality, protein–protein interaction networks

Machine Learning · Deep Learning · AI for Science and Engineering · Bioinformatics

Reach me at [arwang@shanghaitech.edu.cn], or find my code on GitHub.