ByteDance's Anew Labs Presents First AI-Designed Therapy Targeting Previously Undruggable IL-17 Protein
核心洞察
ByteDance (搜索)'s drug discovery unit Anew Labs (搜索) presented its first AI-designed therapy at a major immunology conference in Boston, showcasing a generative AI-designed small molecule targeting IL-17 (搜索), a protein-protein interaction historically considered undruggable.
The company published AnewOmni, a generative AI framework trained on over 5 million biomolecular complexes that claims to be the first to design functional molecules across all molecular scales, from small compounds to peptides to nanobodies.
Anew Labs (搜索) operates with 36 core team members across Shanghai, Singapore, and San Jose, with a scientific advisory board including former executives from Innovent Biologics (搜索), Amgen, and Takeda (搜索), signaling serious pharmaceutical industry ambitions.
ByteDance (搜索)'s drug discovery unit, Anew Labs (搜索), has made its public debut in the pharmaceutical AI space by presenting its first AI-designed therapy at the American Association of Immunologists' annual meeting in Boston. The presentation showcased a generative AI-designed small molecule targeting IL-17 (搜索), a cytokine involved in autoimmune diseases (搜索) including psoriasis (搜索), rheumatoid arthritis (搜索), and ankylosing spondylitis (搜索).
The molecule represents a significant technical achievement as it targets a protein-protein interaction, a category of drug target that the pharmaceutical industry has historically called "undruggable" because the binding surfaces are too large and too flat for conventional small molecules to disrupt. Chris Li, Anew Labs (搜索)' head of biology, presented the pan-spectrum IL-17 (搜索) inhibitor, which is designed to block multiple forms of the IL-17 cytokine rather than a single variant.
Strategic Platform Development
Anew Labs (搜索) has developed AnewOmni, a generative AI framework trained on more than five million biomolecular complexes. According to a preprint published on bioRxiv in March, the model is designed to work across molecular scales, from small chemical compounds to peptides to nanobodies, assembling chemically meaningful building blocks at atomic resolution.
The researchers demonstrated that AnewOmni could design functional molecules targeting KRAS G12D (搜索), one of the most studied oncology targets, and PCSK9, a cholesterol-related protein, achieving success rates between 23 and 75 percent with only low-throughput laboratory validation. The model uses programmable graph prompts that allow researchers to steer the generation process by specifying chemical, geometric, and topological constraints.
Organizational Structure and Expertise
Anew Labs (搜索) operates from offices in Shanghai, Singapore, and San Jose, California, with 36 core team members and a scientific advisory board that includes Liu Yongjun, former president of Innovent Biologics (搜索), Ji Ma, a former principal scientist at Amgen, and Hua Zou, scientific director of protein chemistry at Takeda (搜索) California.
The unit's ambition is to replace expensive injectable antibody therapies with oral pills, using generative AI to design small molecules that can achieve comparable efficacy to existing treatments. Current IL-17 (搜索) therapies, including Novartis's secukinumab and Eli Lilly's ixekizumab, are injectable antibodies that generate billions in annual revenue by treating psoriasis (搜索) and other inflammatory conditions.
Technical Innovation and Competitive Positioning
ByteDance (搜索)'s entry into drug discovery leverages AI infrastructure originally built for TikTok's recommendation engine. The company's generative models process enormous quantities of molecular data and predict which combinations of atoms will produce desired biological responses, applying organizational capabilities developed for serving 1.5 billion TikTok users to molecular binding affinity problems.
The company has also released Protenix-v1, an open-source biomolecular structure prediction model that matches AlphaFold3-level performance in predicting protein, DNA, RNA, and ligand structures. The model code and parameters were made publicly available under an Apache 2.0 license, positioning ByteDance (搜索) as a serious AI research organization in the pharmaceutical space.
Pipeline and Future Prospects
Anew Labs (搜索) currently has four pipeline candidates in development, all focused on oral small molecules for immunology targets, specifically protein-protein interaction targets that have been validated by antibody research but are difficult to drug with conventional approaches. The unit is also developing AnewSampling, a framework designed to reproduce molecular dynamics at the all-atom level to improve simulation accuracy.
The company is expanding its conference presence, with registrations for the BIO International Convention in San Diego and presentations planned for the Free Energy Workshop in Barcelona. This coordinated scientific outreach represents a systematic effort to establish credibility within the pharmaceutical research community.
More than 173 AI-discovered drug programs are currently in clinical development globally, with 15 to 20 entering large-scale trials this year. ByteDance (搜索)'s entry into this competitive landscape brings significant computational resources and a generative platform that, according to its preprint results, can design functional molecules across multiple scales.