ACE

ACE™ (AI Catalyzed Enzyme Engineering) is a GenAI-powered in silico platform developed by Aganitha for enzyme characterization, design, and optimization. Designed for biopharma organizations, ACE™ addresses the challenge of navigating the vast combinatorial space of possible enzyme designs — a task that is impractical with traditional laboratory methods alone. The suite accelerates enzyme engineering by combining proven approaches such as ancestral sequence reconstruction and ML-directed evolution with cutting-edge generative AI solutions.

Enzymes play critical roles in biopharma both as therapeutics and as biocatalysts. ACE™ supports teams working across both of these domains, enabling faster development cycles, reduced wet lab burden, and more rational enzyme design through integrated AI and machine learning workflows.

Therapeutic Enzyme Engineering

• Leverages Ancestral Sequence Reconstruction (ASR) combined with Protein Language Models (PLMs) to identify ancestral enzyme variants that exhibit enhanced thermal stability, substrate flexibility, and activity as robust starting points for therapeutic development.
• Employs deep learning-powered tools to predict key enzyme properties including thermal stability, solubility, and activity, significantly speeding up variant testing and reducing reliance on wet lab experiments.
• Uses deep learning models and Molecular Dynamics (MD) simulations to predict 3D structures and reaction mechanisms, providing insights into enzyme stability and function and guiding selection of ancestral sequences most suitable for laboratory synthesis.

Enzymes as Biocatalysts

• Enhances traditional biocatalysis pipelines through ML-driven directed evolution, using an active learning approach to identify enzyme variants most likely to exhibit improved target traits such as stability, specificity, selectivity, and reaction efficiency — replacing random mutation testing with intelligent, model-guided exploration.
• Employs molecular dynamics simulations for in-depth analysis of enzyme interactions, providing insights into structural changes and reaction mechanisms to make it easier to identify impactful amino acid substitutions.

Enzyme Property Characterization and Optimization

• Integrates multiple AI/ML tools and knowledgebases for both sequence-based and structure-based homolog searches, enabling users to retrieve sequence homologs from sources such as NCBI BLAST, find structural homologs from large protein databases, and conduct multiple sequence and structural alignments.
• Applies deep learning models to predict enzyme properties such as thermal stability and solubility, supporting rational enzyme design and accelerating variant testing.
• Leverages sequence- and structure-aware protein language models to generate evolutionarily favoured enzyme variants, enabling prediction of evolutionarily favourable mutations from large protein databases.

Platform Highlights

• Integrated, no-code workflows: Multiple AI/ML models and data sources are combined into streamlined workflows, reducing manual effort and saving time throughout the discovery process.
• Data privacy and safety: Aganitha brings infrastructure-as-code to the customer's own local or cloud environment, ensuring proprietary data remains secure and under the organization's control.
• Configurable and scalable: The platform supports on-demand cloud-based High Performance Computing (HPC) clusters with workload management techniques, providing scalable computational resources as needed.

ACE™ is part of Aganitha's broader suite of customizable platforms and solutions spanning the drug discovery value chain, from target discovery to therapeutic development. The platform is deployable within a customer's own infrastructure and is designed to complement existing biopharma R&D workflows across therapeutic areas including oncology, rare diseases, immunology, and CNS disorders.