Iambic Therapeutics has unveiled Enchant, a revolutionary multi-modal AI transformer designed to address a critical bottleneck in drug discovery and development: the limited availability of clinical data. With clinical trials often failing in late stages and drug development costs reaching astronomical levels, the industry has struggled to bring new treatments to patients with unmet medical needs. Enchant aims to change that.
Built on cutting-edge AI architecture, Enchant leverages vast quantities of data from the discovery phase—when thousands of molecules are screened in the lab—and bridges the gap to the clinical phase, where a single drug is tested in human volunteers. Traditionally, while abundant lab-generated data has facilitated advancements in target identification, molecular discovery, and trial design, the leap to clinical prediction has been hindered by a scarcity of human data. Economic and ethical constraints limit testing, creating what Iambic calls a “data wall” between preclinical and clinical stages.
Enchant breaks through this data wall. By training on preclinical data and just a small fraction of clinical data, the model generates accurate predictions for clinical outcomes, including crucial human pharmacokinetics (PK) parameters. Demonstrated results show that Enchant can reliably predict human PK outcomes when trained on clinical data for as few as five molecules, less than 1% of the full dataset. In trials with limited clinical data, Enchant’s predictions outperform those of existing state-of-the-art technologies, proving its value even with data scarcity. When fully trained on both preclinical data and the Obach dataset of clinical PK properties, Enchant’s predictive accuracy reaches a Spearman correlation of 0.74 for human PK half-life, far surpassing the prior industry best of 0.58.
As per the press release, this achievement is a testament to Enchant’s unique ability to harness discovery data to predict clinical results, setting a new benchmark for the industry. Beyond achieving predictive power, Enchant has enabled Iambic Therapeutics to mitigate clinical risk from the outset of its programs. The model’s insights help flag potential issues early, making the drug discovery process more efficient and significantly lowering the likelihood of late-stage failures.
Enchant’s successful debut marks a pivotal step forward in AI-driven drug development, promising safer, more effective therapies.