neoag
Neoantigen × AI
Daily research signal
Industry · 2026-10-21

LG Chem Targets 2029 Clinical Trials for AI-Designed Personalized Neoantigen mRNA Cancer Vaccine

LG Chem plans preclinical testing in 2027 and clinical development by 2029 for a personalized mRNA cancer vaccine that uses AI to predict patient-specific neoantigens. The strategy pairs AI-driven biomarker discovery with patient selection to advance precision oncology.

LG Chem has set a 2029 target to begin clinical development of a personalized mRNA cancer vaccine, with preclinical testing planned for 2027. The Korean chemical and pharmaceutical company unveiled the timeline at the inaugural LG AI Bio Conference in Seoul, co-hosted with LG AI Research and US biotech venture capital firm Flagship Pioneering. At the program's core is the use of artificial intelligence to predict patient-specific neoantigens — tumor-specific antigens that can trigger immune responses against cancer cells — delivered in an mRNA vaccine format designed for individual patients.

The distinguishing element of the strategy is not the vaccine modality alone but the pairing of AI-driven neoantigen prediction with AI models that identify biomarkers predicting treatment response and enabling patient selection for clinical trials. LG Chem frames this dual computational approach as the foundation of its precision oncology strategy: the same AI infrastructure that selects which antigens to target also selects which patients to enroll. Son Jee-woong, president and head of LG Chem's life sciences division, summarized the thesis: "Future competitiveness in drug development depends not only on finding the right drug, but also on identifying the right patient through AI."

The announcement lands in a field where computational maturity is increasingly the differentiator. Open bioinformatics workflows for neoantigen prediction and prioritization have consolidated in recent years, and neoantigen-based vaccines are reaching later-stage trials, raising the bar for how companies demonstrate that in-house AI selection engines outperform established pipelines. LG Chem's 2027 and 2029 checkpoints give the field concrete dates against which to judge whether an industrial-chemicals incumbent can convert AI investment into clinical-stage personalized immunotherapy.

AI-Driven Neoantigen Prediction Pipeline

LG Chem's vaccine program uses AI to predict patient-specific neoantigens, the tumor-specific antigens that personalized vaccines harness to stimulate anti-tumor immune responses. The computational problem is substantial: designing such a vaccine requires sophisticated workflows to predict and prioritize neoantigen candidates from patient data, a challenge addressed by recent work such as the ImmunoNX workflow (2025), built as a robust bioinformatics pipeline to support personalized neoantigen vaccine trials.

The tooling landscape against which LG Chem's models will be measured is well defined. pVACtools, an open-source informatic suite first introduced in 2016 and updated through version 6 in 2026, supports neoantigen prediction, visualization, and therapy design; its developers note that the rise of checkpoint blockade therapies and neoantigen-based vaccines reaching later-stage trials has created a growing need for computational tools to identify and prioritize neoantigens. Any proprietary AI predictor entering clinical development will need to demonstrate performance against such established, openly scrutinized baselines.

Computational epitope selection also has a longer lineage than the current neoantigen wave: epitope analysis for influenza vaccine design was being framed statistically as early as 2004, when researchers moved beyond whole-sequence comparisons toward models linking sequence differences to vaccine effectiveness. The neoantigen field inherits this premise — that antigen selection is a prediction problem — and LG Chem is betting that modern AI can solve it at individual-patient resolution.

Biomarkers for Patient Selection

Alongside the vaccine program, LG Chem is developing AI models to identify biomarkers that predict treatment responses and help select patients for clinical trials — a component the company describes as a key part of its precision oncology strategy. In personalized neoantigen vaccination, where each product is built for a single patient, enrollment of patients likely to respond is arguably as consequential as antigen selection itself; a poorly selected cohort can obscure a biologically active vaccine.

The difficulty of the biomarker problem is well documented. Biomarker discovery for multi-factorial diseases has remained challenging, particularly for indicators that signal changes in health status very early, and a 2024 analysis explores whether quantum computing could enhance biomarker discovery where classical approaches fall short. LG Chem's wager is that machine learning over clinical and molecular data can surface response-predictive signatures that conventional single-marker approaches miss, and that these signatures can be operationalized as trial enrollment criteria.

The strategic logic connects the two halves of the program: an AI system trained to predict which neoantigens are immunogenic is, in principle, positioned to learn which patients carry immunogenic antigen landscapes. Whether LG Chem can demonstrate that linkage — predictive models that perform for both antigen selection and patient selection — will determine if the "right drug, right patient" framing translates into trial results.

Korea's AI Bio Convening

The venue for the announcement was as deliberate as the timeline. The inaugural LG AI Bio Conference, co-hosted by LG Chem, LG AI Research, and Flagship Pioneering, positioned the vaccine program within a broader AI-in-drug-development agenda. Andre Andonian, Flagship Pioneering's Asia-Pacific chair, highlighted AI's expanding role in biotechnology, citing Moderna and Lila Sciences as reference points, while Insilico Medicine's Jue Wang presented a generative AI-discovered drug candidate that has advanced into clinical development.

For LG Chem — a chemicals and pharmaceutical company building out a life sciences division under Son Jee-woong — the conference framing signals that AI is being treated as core discovery and development infrastructure rather than a peripheral digitization effort. Aligning with a US biotech venture firm and showcasing third-party AI drug development successes suggests an intent to connect Korean industrial scale to global AI-biotech networks, with the personalized cancer vaccine as the flagship proof point.

What to Watch

The 2027 preclinical start is the first hard checkpoint. Watch for LG Chem to disclose the architecture and validation data behind its neoantigen prediction models — benchmark performance against established open workflows of the pVACtools and ImmunoNX type would be the clearest evidence that the in-house AI offers more than repackaged standard practice.

The biomarker and enrollment strategy deserves equal scrutiny. Which tumor indications LG Chem selects, which response-predictive biomarkers its models nominate, and how those biomarkers are written into clinical trial eligibility criteria will show whether the "identifying the right patient through AI" thesis is operational or rhetorical. Any disclosed partnerships for sequencing, clinical data access, or manufacturing would further indicate how the 2029 clinical timeline is resourced.

Field-level signals matter as well: as neoantigen-based vaccines advance in later-stage trials elsewhere and open computational suites continue to mature, the competitive baseline for LG Chem's entry keeps rising. Successive editions of the LG AI Bio Conference and the depth of the Flagship Pioneering relationship will indicate whether the 2029 program is a one-off announcement or the leading edge of a sustained AI-driven oncology effort.

Sources

mRNA vaccines

Explore the field

TodayTopicsTrials