Research presented at ISMB 2026 highlights AI's expanding role in drug discovery
Artificial intelligence is rapidly evolving from a research support tool into a core engine for discovering and designing new medicines. As pharmaceutical companies race to integrate AI into drug development, Hanmi Pharm has demonstrated how proprietary AI technology can generate novel drug candidates and accelerate them into clinical development.
Hanmi Pharm announced on July 24 that it presented research at the International Conference on Intelligent Systems for Molecular Biology (ISMB 2026), held in Washington, D.C., showcasing its proprietary AI platform, HARP-pSAR (Hanmi AI-driven Research Platform – protein Sequence-Activity Relationship).
The research demonstrated that AI can move beyond predicting protein activity to designing optimized protein sequences that ultimately become clinical-stage drug candidates.
The company's flagship pipeline generated through the platform is HM17321 (development code: LA-UCN2), a first-in-class obesity treatment designed to simultaneously reduce body fat while increasing skeletal muscle mass. The candidate is currently undergoing Phase 1 clinical trials in the United States.
Unlike conventional obesity therapies that primarily focus on minimizing muscle loss during weight reduction, HM17321 aims to selectively reduce fat while actively promoting muscle growth, representing a fundamentally new therapeutic approach.
HM17321 is designed as a Urocortin-2 (UCN2) analog that selectively targets the corticotropin-releasing factor 2 (CRF2) receptor rather than incretin receptors such as GLP-1. According to Hanmi Pharm, selective activation of the CRF2 receptor can directly promote fat reduction, muscle growth, and improved muscle function.
One of the greatest scientific challenges was designing a UCN2 analog that binds strongly to the target CRF2 receptor while avoiding activation of the non-target CRF1 receptor, which could lead to undesirable biological effects.
To address this challenge, Hanmi Pharm employed its proprietary HARP-pSAR platform. The AI model predicts the biological activity of newly designed protein sequences using only a limited amount of experimental data and recommends optimal amino acid substitutions for candidate optimization.
Unlike many AI models that require massive datasets and high-performance computing infrastructure, HARP-pSAR was developed to achieve high predictive accuracy even with relatively small experimental datasets. By incorporating quantization technology, the platform significantly reduces computational costs while remaining capable of operating efficiently on standard research workstations.
Using AI-guided screening, researchers were able to rapidly identify promising candidates instead of synthesizing and experimentally evaluating countless protein variants through conventional trial-and-error methods.
The platform also predicts the activity of entirely new protein sequences that were absent from the training dataset and recommends optimal positions for novel amino acid residues. This capability substantially reduces experimental costs and development time while improving the overall quality of candidate optimization.
Hanmi Pharm noted that this iterative process—where AI-generated designs are experimentally validated and the results are fed back into the learning model—closely aligns with the emerging concept of AI-native drug discovery that is gaining momentum across the global pharmaceutical industry.
Beyond obesity therapeutics, the company is also expanding its muscle-enhancing pipeline. At the American Diabetes Association (ADA) Scientific Sessions earlier this year, Hanmi Pharm introduced HM500197 (LA-MSTN), described as the world's first peptide-based myostatin inhibitor designed to enhance skeletal muscle growth.
ISMB is widely recognized as one of the world's leading conferences in bioinformatics and computational biology. This year's meeting placed particular emphasis on AI systems capable of solving real-world research challenges and producing tangible drug discovery outcomes, rather than merely demonstrating predictive performance.
Choi In-young, Vice President and Head of Future Growth at Hanmi Pharm, said, "This research demonstrates that AI can serve not only as a data analysis tool but also as a true research partner that collaborates with scientists in designing innovative drug candidates and reducing trial-and-error throughout the drug discovery process."
She added, "Presenting at ISMB, where global leaders in AI and computational biology gather, was particularly meaningful because we were able to demonstrate that AI-designed candidates can successfully advance into clinical drug development. We plan to further expand the HARP-pSAR platform into a broad range of therapeutic areas."
