The Next Frontier in Medicine
In the intricate dance of life, proteins have long stolen the spotlight in drug development. For decades, pharmaceutical companies have focused almost exclusively on targeting these molecular workhorses. Yet, there's a crucial flaw in this approach – many disease-causing proteins are considered "undruggable" due to their flat surfaces or complex structures. Meanwhile, hidden in plain sight lies a more fundamental player: RNA, the versatile molecule that translates genetic blueprints into reality.
At Tsinghua University, a revolution is quietly unfolding. Scientists are pioneering methods to target RNA directly – the very computational engine of the cell. Their work represents a paradigm shift in therapeutic development, potentially opening up thousands of new targets for diseases that have long evaded treatment.
By combining cutting-edge artificial intelligence with deep biological insight, these researchers are not just discovering new drugs; they're rewriting the rules of how we intervene in disease processes.
of human genome codes for proteins
of genome transcribed into RNA
of genome targetable with traditional approaches
The case for targeting RNA is both simple and compelling. While only about 1.5% of the human genome actually codes for proteins, approximately 70% of our genome is transcribed into RNA. This includes vast stretches of non-coding RNA (ncRNA) that play crucial regulatory roles in health and disease 2 .
Consider the numbers: of the roughly 20,000 protein-coding genes in humans, only an estimated 700-900 produce proteins considered "druggable" with current technologies. This represents a mere 0.05% of the total length of the human genome that we can effectively target with traditional approaches 2 . The remaining untapped territory – including disease-linked non-coding RNAs – represents an immense opportunity for therapeutic development.
"The origin of life is an RNA world," observes Professor Zhi John Lu's team at Tsinghua's School of Life Sciences. "This seemingly simple RNA language may have the basic elements of the origin of life or even the origin of the universe: information replication, transmission and mutation" 2 .
Human Genome Distribution
The greatest challenge in RNA-targeted drug development has been the lack of high-resolution RNA structures. Traditional drug design methods rely heavily on detailed 3D structural information to identify binding pockets and predict molecular interactions. For most disease-relevant RNA molecules, this structural information simply doesn't exist.
In July 2025, Professor Zhi John Lu's team unveiled a groundbreaking solution: RNAsmol, an artificial intelligence model that predicts how small molecules interact with RNA targets using only sequence information 2 . This breakthrough circumvents the structural barrier entirely, opening the floodgates for targeting countless RNA molecules previously considered "undruggable."
The system randomly perturbs training data to simulate real-world diversity, helping the model better learn the rules of RNA-small molecule binding 2 .
By generating virtual negative samples and potential unlabeled samples based on known interactions, the model expands its understanding of the interaction landscape 2 .
The model combines graph-based molecular feature representation with a feature fusion module that uses attention mechanisms to weightedly integrate target and drug molecule features 2 .
The results are striking. RNAsmol outperforms traditional methods by approximately 8% in average AUROC (area under the curve) in 10-fold cross-validation, with performance improvements of about 16% in evaluating unseen samples 2 . In practical applications, it improved ranking scores by roughly 30% when distinguishing between bait molecules and real ligands – a critical capability for virtual screening in drug discovery 2 .
| Metric | Traditional Methods | RNAsmol | Improvement |
|---|---|---|---|
| Average AUROC (10-fold cross-validation) | Baseline | +8% | 8% |
| Performance on unseen samples | Baseline | +16% | 16% |
| Virtual screening ranking score | Baseline | +30% | 30% |
While RNA-targeted therapies represent one frontier, Tsinghua's life sciences research spans a much broader landscape. The university has established a comprehensive research framework focusing on five key pillars: small molecules, biomacromolecules, genes, cell therapy, and medical devices 4 . This strategic approach breaks through traditional subject classifications to emphasize distinctive characteristics and interdisciplinary development.
At the Tsinghua-PKU Joint Center for Life Sciences (CLS), researchers have produced substantial output, with 278 articles accounting for 56.60 Share in the Nature Index during the 12-month period from August 2024 to July 2025 6 . The biological sciences dominate this output (44.59 Share), followed by chemistry (21.02 Share) 6 .
RNAsmol AI model, circular mRNA vectors for treatment of genetic disorders, cancer, infectious diseases.
Artificial cells, self-amplifying mRNA for programmable therapeutics, sustainable manufacturing.
Photocatalytic plastic degradation, CO₂ capture for pollution control, climate change mitigation.
Microrobot swarms, conducting polymer hydrogels for targeted drug delivery, bioelectronic medicines.
| Research Field | Example Projects | Potential Applications |
|---|---|---|
| RNA-Targeted Therapeutics | RNAsmol AI model, circular mRNA vectors | Treatment of genetic disorders, cancer, infectious diseases |
| Synthetic Biology | Artificial cells, self-amplifying mRNA | Programmable therapeutics, sustainable manufacturing |
| Environmental Biotechnology | Photocatalytic plastic degradation, CO₂ capture | Pollution control, climate change mitigation |
| Nanobiotechnology | Microrobot swarms, conducting polymer hydrogels | Targeted drug delivery, bioelectronic medicines |
Cutting-edge research requires advanced tools. At Tsinghua, scientists have access to a comprehensive suite of technologies that enable their pioneering work:
Including PCR, qPCR, sequencing, microarrays, and molecular biology tools 3 .
Featuring antibodies, ELISA, protein production, western blots, and protein microarrays 3 .
Encompassing discovery services, microscopy, flow cytometry, transfection, and cell culture systems 3 .
Including liquid chromatography, mass spectroscopy, structural analysis methods, and spectroscopy 3 .
This technological ecosystem supports the entire research pipeline – from basic discovery to therapeutic development. The School of Pharmaceutical Sciences further strengthens this pipeline by focusing on three general directions: key biological issues of pharmacy, key technology research of pharmacy, and disease research and targeted therapy 4 .
| Technology Category | Key Tools and Reagents | Research Functions |
|---|---|---|
| Nucleic Acid Analysis | qPCR instruments, sequencing platforms, purification kits | Gene expression analysis, genome sequencing, sample preparation |
| Protein Characterization | Antibodies, mass spectrometers, electrophoresis systems | Protein identification, quantification, and functional analysis |
| Cell Culture and Analysis | Cell media, transfection reagents, flow cytometers | Cell maintenance, genetic modification, cell population analysis |
| Data Analysis and Management | Bioinformatics software, LIMS, computational models | Data processing, storage, and knowledge extraction |
Despite exciting progress, significant challenges remain in RNA-targeted drug development. Delivery efficiency, tissue specificity, and potential off-target effects represent hurdles that Tsinghua researchers continue to address through innovative delivery systems and enhanced prediction algorithms.
The future direction of life sciences research at Tsinghua appears poised for further expansion. The School of Pharmaceutical Sciences plans to grow its principal investigators to 60 researchers over the next decade, forming outstanding research echelons led by internationally renowned scholars and driven by excellent young and middle-aged scientists 4 .
International collaboration will play a crucial role in this expansion. Currently, the Tsinghua-PKU Joint Center for Life Sciences maintains 13.5% international collaboration in its research output, with leading partners including Harvard University, Stanford University, and the University of California 6 .
International Collaboration
The work underway at Tsinghua University represents more than incremental advances – it signals a fundamental shift in how we approach disease treatment. By targeting RNA rather than just proteins, scientists are intervening earlier in the disease process, potentially addressing conditions that have long resisted conventional therapies.
The integration of artificial intelligence with biological discovery, as exemplified by the RNAsmol platform, demonstrates how computational approaches can overcome longstanding barriers in drug development. Meanwhile, the broader synthetic biology efforts underway at Tsinghua highlight how engineering principles can be applied to biological systems, creating not just new medicines but potentially new forms of sustainable manufacturing and environmental remediation.
As Professor Lu's team reflects, "RNA, not DNA, is the computational engine of the cell" 2 . By learning to program this fundamental machinery, Tsinghua scientists are not just developing new drugs – they're advancing toward a future where we can truly reprogram biology to promote human health and environmental sustainability.