AI-Powered In Silico Modeling for Biologics

Compare up to 16 candidates based on sequence and structural analysis, developability liabilities, and benchmarking against clinical-stage therapeutic proteins.

Best suited for: Lead optimization and candidate selection

Typical outputs: Developability assessment, therapeutic-protein benchmarking, comparative candidate ranking, progression recommendations, and expert scientific report.

Assess up to eight candidates for self-association, viscosity risk, and protein-protein interactions using molecular dynamics simulations.

Best suited for: Candidate selection for high-concentration products

Typical outputs: High-concentration suitability assessment, viscosity and self-association risk evaluation, molecular interaction insights, candidate ranking, and expert recommendations.

Compare up to four candidates across selected pH and ionic-strength conditions to identify formulation sensitivities and assess candidate robustness.

Best suited for: Late candidate selection and early formulation decision-making

Typical outputs: Formulation sensitivity profiles, robustness assessment across formulation conditions, comparative candidate ranking, and formulation development recommendations.

Explore the formulation space of a selected lead candidate to identify promising pH, ionic-strength, and functional excipient regions for experimental development.

Best suited for: Preformulation and formulation-development planning

Typical outputs: Prioritized formulation corridors, excipient-class assessment, formulation strategy recommendations, and an experimental screening plan.

Yes. Coriolis can start an in silico assessment using only the primary amino acid sequence of your biologic candidate.

Our AI and data-driven models and mechanistic molecular dynamics simulations evaluate key developability risks, protein stability, formulation challenges, and molecular interactions before laboratory material is available. This enables earlier decision-making, faster candidate selection, and more focused experimental studies later in development.

In silico modeling is a powerful decision-support tool that helps identify developability risks, compare candidates, and prioritize formulation strategies early in development. It can provide robust relative metrics for comparing colloidal and conformational stability without requiring prediction of the absolute value of a final critical quality attribute (CQA).

While computational predictions do not replace stability-indicating measurements of CQAs, they can significantly reduce uncertainty and improve candidate-selection decisions.

Coriolis combines AI-based prediction, molecular dynamics simulations, and high-throughput biophysical characterization to generate actionable insights that support more efficient biologics development.

In silico benchmarking helps evaluate how a biologic candidate compares to clinically advanced therapeutic proteins, providing valuable context for candidate selection and formulation development.

Coriolis compares calculated biophysical descriptors against more than 700 clinical-stage therapeutic proteins, including 600+ mAbs and 100+ non-mAb biologics. For mAb-like molecules, comparison against 80+ marketed formulations can further support formulation strategy decisions for candidates with similar physicochemical profiles.

Modern molecular dynamics simulations can deliver valuable insights early in development while reducing costly experimental iterations. Coriolis applies both coarse-grained and atomistic molecular dynamics approaches, selecting the most efficient model based on molecule type and development goals.

For larger proteins such as mAbs, our in-house coarse-grained approach can reduce computational effort by up to 25-fold compared with full-atom simulations, enabling efficient assessment of formulation conditions such as pH, ionic strength, and selected excipient classes. Full-atom models can additionally account for non-canonical amino acids, lipidation, glycosylation, conjugated moieties, and selected excipients.

This allows us to assess self-interaction, formulation sensitivity, and excipient effects while maintaining practical timelines and budgets.

In silico modeling is especially valuable for novel biologics where limited experimental data exists. Coriolis combines de novo structure prediction, homology modeling, mechanistic molecular dynamics simulations, and formulation expertise to evaluate complex proteins, peptides, antibody fragments, and modified biologics.

For highly flexible molecules such as peptides, full-atom molecular dynamics can sample multiple conformations and identify representative conformational clusters.

While machine-learning models depend on the availability of representative training data, mechanistic molecular dynamics simulations do not require modality-specific datasets. This approach helps identify developability risks, evaluate conformation and self-interaction, assess formulation options, and guide candidate selection even when historical datasets are limited.

Glycosylation can significantly influence protein stability, molecular interactions, and formulation behavior.

Coriolis evaluates glycosylation patterns, including glycan location, degree of sialylation, and glycoform distribution, and incorporates them into the appropriate in silico modeling strategy.

Depending on the molecule and its complexity, we apply coarse-grained or full-atom molecular dynamics simulations to assess how glycans may affect conformational and colloidal stability, molecular interactions, developability and formulation performance.

Where a representative glycoform can be defined, it can be explicitly incorporated into full-atom simulations to support more informed biologics development decisions.