NTU Singapore Spin-off

AI-Powered Drug Target Discovery from Single-Cell Data

Cytorion Biotech harnesses GATomics — a Graph Attention Network engine — to identify novel therapeutic targets across disease progression, accelerating the path from multi-omics insight to validated drug candidates.

Our Pipeline → Explore GATomics
19
Disease Datasets Validated
13/19
1st Place Rankings
9
Cancer Types Tested
About Us

Bridging AI and
Drug Discovery

Cytorion Biotech is a Singapore-based AI drug discovery company spun out of Nanyang Technological University (NTU). We are pioneering a new approach to understanding disease progression at the single-cell level.

Traditional methods treat disease stages as discrete categories with clear boundaries — but biology is continuous. Our proprietary GATomics platform uses Graph Attention Networks to capture the global, dynamic changes across multiple disease stages simultaneously, identifying drug targets that conventional tools miss.

From novel target identification to AI-based virtual screening and wet-lab pharmacological validation, we operate a fully integrated discovery pipeline that has already yielded promising lead compounds across multiple cancer indications.

🧬 Single-Cell Resolution

Cell-level analysis captures disease heterogeneity that bulk methods overlook, enabling precise identification of progression-related genes.

🔄 Continuous Disease Modeling

Unlike discrete methods, GATomics models disease as a continuous spectrum, capturing subtle transitions across multiple stages.

🎯 Target to Compound

End-to-end pipeline from AI target discovery through virtual screening (10M+ molecules) to pharmacological validation in wet lab.

🏥 Multi-Cancer Applicability

Validated across AML, Multiple Myeloma, NSCLC, PDAC and more — demonstrating broad therapeutic potential.

Core Technology

The GATomics Platform

A Graph Attention Network-based AI engine for multi-class differentially expressed gene analysis at single-cell resolution — going beyond static, binary comparisons.

Step 01

Multi-Omics Input

Ingest single-cell RNA-seq and ATAC-seq data across multiple disease stages (e.g. Diagnosis → Remission → Relapse).

→
Step 02

Graph Construction

Build cell-gene interaction graphs that preserve cell-level information and multi-stage relationships simultaneously.

→
Step 03

Attention-Based Inference

Graph Attention Networks learn which genes and cell interactions drive disease progression across all stages globally.

→
Step 04

Target Prioritization

Output ranked gene lists validated against KEGG, GO, Reactome, DisGeNet, and DepMap functional databases.

13×

GATomics ranked 1st place in 13 out of 19 disease datasets, outperforming Scanpy, ANOVA, and Random Forest.

223%

Average precision improvement over Scanpy in mapping known disease driver genes across 19 datasets.

5/8

Won 1st place in 5 out of 8 cancer types in DepMap CRISPR/RNAi KO/KD functional validation benchmarks.

10M+

Molecules screened through our AI virtual screening funnel — from BIND to Smina docking to GROMACS MD simulation.

Drug Pipeline

Therapeutic Programs

Our lead programs target novel biology identified by GATomics, with validated pharmacological activity across multiple cancer types and confirmed target binding.

Program / Target Indication Discovery Hit to Lead Preclinical Status
D8 → Gene X AML (MLL-AF9, AML1-ETO)
Lead Optimization
D8 → Gene X Multiple Myeloma (Dex-sensitive)
Hit Validation
D8 → Gene X NSCLC (p53-null)
Hit Validation
D8 → Gene X PDAC (KRAS G12D)
Hit Validation

D8 → Gene X

AML (MLL-AF9, AML1-ETO)
Lead OptimizationActive

D8 → Gene X

Multiple Myeloma (Dex-sensitive)
Hit ValidationActive

D8 → Gene X

NSCLC (p53-null)
Hit ValidationActive

D8 → Gene X

PDAC (KRAS G12D)
Hit ValidationActive

Key Findings: D8 shows efficacy against 5 leukemia cell lines with IC₅₀ as low as 4.56 μM (THP-1, D8RT). ITC confirmed direct binding to Protein X (KD = 513 nM). Compound shows activity independent of TP53 status and RAS alterations. MYC amplification is the only genomic feature associated with reduced efficacy. Normal bone marrow cells show no biological mechanism damage — indicating a clear therapeutic window.

Our Team

World-Class Expertise

A multidisciplinary team combining deep expertise in computational biology, drug discovery, and business strategy.

Prof. Mu Yuguang

Prof. Mu Yuguang

Team Leader

Associate Professor at NTU, leading the Biomolecular Simulations and Data Analysis Lab. 30+ years in computer-aided drug design with 200+ SCI publications in high-impact journals.

Sun Ao

Sun Ao

Co-Founder

PhD researcher in Dr. Mu's lab and 1st author of GATomics. 7 years of wet-lab experience in drug discovery and immunology, with expertise in pharmacology, immunology, and deep learning.

Zhang Zhenlin

Zhang Zhenlin

Co-Founder

MBA from Shanghai Jiao Tong University. 12 years of executive experience in Hong Kong-listed companies. 8 years as CEO of an investment firm, expert in corporate management and financing.

Get in Touch

Partner with Us

Interested in our AI drug discovery platform or pipeline? We welcome partnership, licensing, and investment inquiries.

📍

Location

School of Biological Sciences
Nanyang Technological University
Singapore

✉️
🌐