Tevogen Marks Major Milestone in Its AI Initiative to Enhance Efficacy of T Cell–Based Therapies with 100x PredicTcell Beta Data Expansion

• Beta version of PredicTcell™ expands training dataset to ~1.4 million and total dataset to over 6.7 billion records.

• Potential future T cell therapies could reliably bind to their target nearly every time, dramatically raising the probability of success.

Tevogen announced significant progress in the development of its proprietary PredicTcell™ platform, designed to accelerate precision immunotherapy development and efficacy through advanced machine learning and transformer-based models. The platform is being developed with the support of Microsoft (Nasdaq: MSFT) and Databricks, leveraging their advanced cloud and data technologies to enable scalability and efficiency.

The alpha version of PredicTcell was trained on more than 124,000 records using transformer-based architecture and 91,000 records using traditional machine learning architecture. The alpha version of the model delivered recall levels of ~92–97% and a precision range of ~38–43%, serving as a proof-of-concept for AI-driven prediction of immunologically active peptide-T cell receptor interactions.

Also Read: AiThority Interview with Tim Morrs, CEO at SpeakUp

Informed by insights from Tevogen’s proprietary ExacTcell™ platform and positive Phase 1 trial results, PredicTcell Beta aims for significantly higher precision in identifying virology targets. Key advancements include:

• Expanded the training dataset tenfold to approximately 1.4 million records, while the total dataset has grown more than 100-fold to over 6.7 billion records.
• Analyzed more than 10.7 billion data points to construct the training set consisting of 6.5 billion virology datapoints, 4.2 billion genomic datapoints, and 416 million oncology datapoints.
• Expanded the number of features for training the model from 22 to 27.

By dramatically scaling its data pipelines and fine-tuning its models, Tevogen.AI could move toward an unprecedented position: T-cell therapies that could accurately bind to their intended target nearly every time. Such predictability would be a transformative breakthrough in medicine, with far-reaching implications:

• Clinical success rates dramatically increased.
• Development timelines shortened, reducing drug costs.
• Greater patient access to life-saving therapies across infectious diseases, oncology, and beyond.

“The promise of PredicTcell goes far beyond data,” said Mittul Mehta, CIO and Head of Tevogen.AI. “If our tools continue to deliver as they have so far, Tevogen stands to create T-cell therapies where binding to the target virus or disease isn’t just probable, but nearly guaranteed. That would mean clinical success, faster c****, reduced costs, and ultimately more lives saved.”

“With the right skill and the proper blend of AI and biotechnology, we can scale discoveries in precision medicine once thought impossible. Tevogen’s AI initiative is to raise efficacy standards in T cell–based therapies, cut development costs, and unlock entirely new markets in immunotherapy,” added Ryan Saadi, M.D., M.P.H., Chief Executive Officer of Tevogen Bio.

Also Read: Cognitive Product Design: Empowering Non-Technical Users Through Natural Language Interaction With AI-Native PLM