Advancing Large-scale in silico Functional Genomics with TEA-GCN
28 July 2026
Scientific Breakthrough
Understanding how genes work together remains one of the central challenges in modern biology. While individual genes have been extensively studied, their interactions within complex biological networks are often difficult to identify and characterize.
To address this challenge, Peng Ken Lim, a PhD student from NTU’s School of Biological Sciences (SBS), led the development of TEA-GCN, an innovative AI-enhanced computational framework designed to infer gene co-expression networks from large-scale transcriptomics data. The tool analyzes patterns of gene activity across diverse biological conditions to predict how genes interact, regulate one another, and function within specific tissues and environments.
To demonstrate its robustness and generalizability, the research team benchmarked TEA-GCN against existing state-of-the-art methods using an unprecedented dataset of more than 450,000 RNA-sequencing (RNA-seq) samples spanning 12 species. The study showed that TEA-GCN consistently delivered superior performance, providing researchers with a more accurate tool for discovering gene functions, studying gene regulation, and comparing biological systems across species.
The significance of the work was recognised through its publication in Nature Communications, reflecting both the scientific rigor of the study and the impact of the methodology on the broader functional genomics research community.
How SingAREN and HPCC Made It Possible
The project was accomplished by a remarkably small team of just four early-career researchers, comprising doctoral and undergraduate students from Associate Professor Marek Mutwil’s Lab, with Peng Ken Lim serving as lead researcher.
A major challenge was the sheer scale of the validation effort. Downloading, managing, and analysing more than 450,000 RNA-seq datasets from international repositories required both high-speed connectivity and substantial computational resources.
SingAREN’s high-performance research network played a critical enabling role by providing fast and reliable access to vast amounts of genomic data hosted worldwide. The ability to transfer such large datasets efficiently allowed the team to pursue a level of benchmarking that would otherwise have been prohibitively time-consuming and resource-intensive.
At the same time, NTU’s High Performance Computing Centre (HPCC) at Nanyang Technological University provided the computational infrastructure needed to process and analyse the massive volume of sequencing data. HPCC’s advanced computing capabilities enabled the large-scale workflows required to optimize, benchmark, and validate TEA-GCN across hundreds of thousands of samples, turning an ambitious research vision into a practical reality.
Impact
By combining SingAREN’s high-speed research connectivity with HPCC’s powerful computing resources, the team was able to conduct one of the most comprehensive evaluations of a gene regulatory network prediction method to date. This integrated digital research infrastructure empowered a small, highly innovative team led by Peng Ken Lim to deliver world-class scientific results, advancing the field of computational genomics and contributing a valuable new tool to researchers worldwide.
(Note: Published in Nature Communications, https://www.nature.com/