Key Takeaways
- •On June 17, 2026, LG AI Research and D&D Pharmatech revealed a new partnership aimed at accelerating the discovery of oral peptide drugs.
- •The partnership is structured around a joint research effort.
- •Peptides occupy an unusual space in drug discovery.
Collaboration Announced
On June 17, 2026, LG AI Research and D&D Pharmatech revealed a new partnership aimed at accelerating the discovery of oral peptide drugs. The announcement, reported by Yahoo Finance, brings together two distinct areas of expertise: LG AI Research’s artificial intelligence and machine learning capabilities and D&D Pharmatech’s specialized knowledge in peptide therapeutics. Their shared goal is to overcome longstanding barriers that have limited peptides to injectable forms and to open the door to orally administered treatments.
The partnership is structured around a joint research effort. LG AI Research will contribute its advanced AI models for molecular design and property prediction. D&D Pharmatech will provide its proprietary platform for discovering and developing peptide-based therapeutics, including experience with cyclic peptides and other formats that may improve stability and permeability. While the specific financial terms of the collaboration were not disclosed, the announcement signals a growing trend among biopharmaceutical companies to combine computational power with biological expertise in pursuit of difficult drug modalities.
The Challenge of Oral Bioavailability
Peptides occupy an unusual space in drug discovery. They are often highly potent and selective, making them attractive for targeting protein protein interactions that small molecules cannot easily reach. Yet they possess inherent liabilities when taken orally. The human gastrointestinal tract is designed to break down proteins and peptides. Enzymes such as pepsin, trypsin, and chymotrypsin degrade peptides into amino acids and short fragments long before they can reach the bloodstream. Even if a peptide survives enzymatic attack, its size and polarity usually prevent it from crossing the intestinal wall. This combination of poor metabolic stability and low permeability means that virtually all peptide drugs in clinical use today must be injected.
Researchers have spent decades trying to overcome these barriers. Strategies include chemical modifications to resist proteolysis, the use of permeation enhancers that temporarily open tight junctions between intestinal cells, and the design of cyclic or stapled peptides that adopt more rigid conformations. Some oral peptide drugs have reached the market. The most notable is semaglutide (Rybelsus), a glucagon-like peptide 1 (GLP‑1) receptor agonist approved for type 2 diabetes. But Rybelsus requires a large dose (7–14 mg compared with 0.5–1 mg for the injectable version) and must be taken on an empty stomach with minimal water. The bioavailability of oral semaglutide is still only about 1 percent. This illustrates the magnitude of the challenge. An AI driven approach could systematically search for modifications that improve these properties without sacrificing potency.
How AI Tackles Peptide Drug Design
LG AI Research applies deep learning models to molecular tasks such as predicting how a peptide will interact with its target, estimating its permeability across cell membranes, and forecasting its stability in the gut. These models can screen millions of peptide variants in silico, ranking them by likelihood of success before any compound is synthesized. In traditional drug discovery, chemists rely on empirical rules and iterative cycles of synthesis and testing. That process can take months or years for a single series. AI can reduce the cycle time and widen the search space.
The algorithms used in this collaboration are not generic off the shelf tools. LG AI Research has previously developed domain specific models for scientific problems, including protein structure prediction and drug target interaction mapping. Their approach likely involves training on large datasets of peptide sequences and their measured properties, as well as incorporating physical simulations such as molecular dynamics. By understanding the relationship between chemical structure and oral bioavailability, the AI can propose non obvious modifications. For example, it might suggest incorporating non natural amino acids, adding lipophilic side chains, or introducing constraints that stabilize the conformation needed for membrane crossing.
D&D Pharmatech complements this computational effort with experimental validation. The company’s platform includes methods for synthesizing and testing peptide candidates in assays that mimic gastrointestinal conditions. This closed loop design build test learn cycle is essential for AI guided drug discovery. The models generate hypotheses, the biologists test them, and the results feed back into the models to improve predictions. Over time, the system becomes more accurate for the specific class of molecules under investigation.
Broader Implications and Next Steps
The collaboration between LG AI Research and D&D Pharmatech is part of a wider movement to apply artificial intelligence to drug modalities beyond small molecules. Peptides, antibodies, and nucleic acid based drugs have traditionally received less attention from computational chemists because of their size and complexity. However, recent advances in deep learning, particularly transformer models and graph neural networks, have made it feasible to design large molecules with desired properties.
If the partnership succeeds, the direct benefit would be oral formulations for diseases that currently require injected peptide therapies. These include metabolic disorders, diabetes, obesity, certain cancers, and inflammatory conditions. Oral delivery would improve patient adherence and reduce the burden on healthcare systems. Patients would no longer need to store refrigerated injectables, undergo training for self injection, or risk needle phobia.
Beyond the immediate target of oral bioavailability, the AI models developed here could also be applied to other aspects of peptide development, such as optimizing half life, reducing immunogenicity, and improving manufacturability. The platform could be scaled to multiple projects simultaneously, potentially accelerating the entire peptide discovery pipeline.
Neither partner has disclosed specific disease targets or timelines for candidate nomination. At this stage, the announcement is a statement of intent and resource commitment. The field will watch for peer reviewed publications, patent filings, and entry into preclinical development as milestones of progress. The partnership highlights a growing recognition that solving oral peptide delivery requires both cutting edge computation and deep biological insight. Neither alone is sufficient.
Frequently Asked Questions
Q: Why are oral peptide drugs so difficult to develop?
A: Peptides are vulnerable to two major barriers in the gastrointestinal tract. They are rapidly degraded by digestive enzymes, and their large size and polarity prevent passive diffusion across the intestinal lining. Even when modified, most oral peptides achieve only very low bioavailability, often less than 5 percent, which means large doses are required.
Q: How does artificial intelligence help design better oral peptides?
A: AI models can predict how chemical modifications affect properties such as enzymatic stability, membrane permeability, and target binding. They can screen millions of virtual peptide variants in hours rather than months, highlighting the most promising candidates for synthesis and testing. The models improve over time as they learn from experimental results.
Q: What are the specific contributions of LG AI Research and D&D Pharmatech?
A: LG AI Research provides the machine learning algorithms and computational infrastructure for molecular design and prediction. D&D Pharmatech contributes its peptide therapeutics platform, including expertise in peptide chemistry, assay development, and preclinical evaluation. Their combined effort creates a design testing cycle where computational hypotheses are rapidly validated in the lab.
Q: What is the timeline for seeing results from this collaboration?
A: No specific timeline was disclosed in the announcement. Early indicators of progress will include scientific publications, patent applications, and announcements of candidate selection for preclinical development. Given the complexity of oral peptide discovery, it is realistic to expect that several years of research and optimization will be needed before a compound enters clinical trials.
