CD ComputaBio integrates physics-based simulation, AI-driven property prediction, and targeted wet-lab validation to assess peptide developability early. We model aggregation, solubility, proteolytic stability, membrane permeability, immunogenicity, and formulation liabilities—so you can triage candidates before costly synthesis and in vivo studies.
We combine sequence-based amyloid motif scanning, molecular dynamics of self-association, and surface hydrophobicity mapping to predict aggregation hotspots and concentration-dependent behavior.
We estimate intrinsic solubility from 3D descriptors and hydrophobic surface area, then model formulation-dependent solubility using counterion and excipient effects.
We map protease cleavage sites, predict deamidation/isomerization/oxidation liabilities, and model backbone dynamics to identify flexible regions prone to degradation.
We simulate passive diffusion across lipid bilayers, estimate permeability coefficients, and evaluate transporter recognition for oral or cell-penetrating peptide design.
We predict peptide binding to common MHC class I and II alleles, identify T-cell epitope hotspots, and flag sequence motifs associated with anti-drug antibody responses.
We model peptide-excipient interactions, predict viscosity at high concentration, and suggest stabilizing formulation conditions to reduce aggregation and degradation.
We use AMBER and GROMACS with modern force fields (ff19SB, CHARMM36m) to simulate peptide self-association, membrane interactions, and conformational flexibility. Free energy calculations (MM/PBSA, umbrella sampling) quantify aggregation and insertion energetics.
Our in-house models are trained on peptide developability datasets and public biophysical data. We predict solubility, permeability, stability, and immunogenicity from sequence and 3D structure, with uncertainty quantification for triage.
Selected computational predictions are confirmed with rapid orthogonal assays: ThT fluorescence for aggregation, DLS for particle size, RP-HPLC for stability, and PAMPA for permeability. This feedback loop improves model confidence for your series.
We collect your peptide sequences (natural or modified), target context, and intended route of administration. Initial sequence-based scans flag obvious aggregation, deamidation, oxidation, and protease motifs.
If no experimental structure exists, we build 3D models via AlphaFold2 or homology modeling, then generate conformational ensembles with MD to capture flexibility relevant to developability.
Each candidate is profiled across aggregation, solubility, stability, permeability, and immunogenicity using our integrated computational pipeline, producing quantitative scores and mechanistic flags.
We synthesize or procure selected peptides and run rapid biophysical assays to confirm key computational predictions, anchoring the assessment with experimental data.
You receive a comparative report ranking candidates by developability, highlighting liability hotspots, and suggesting specific sequence or formulation changes to rescue promising candidates.
| Challenge | Computational Approach | Key Output | Decision Enabled |
|---|---|---|---|
| Peptide aggregates at high concentration | MD self-association + surface hydrophobicity mapping | Aggregation hotspot residues, dimerization free energy | Redesign specific positions to reduce self-association |
| Poor aqueous solubility limits dosing | 3D descriptor-based solubility prediction + charge mapping | Predicted solubility (µg/mL), pH-dependent profile | Select analogs with improved solubility or formulate at suitable pH |
| Rapid serum degradation | Protease cleavage site prediction + MD flexibility | Cleavage site ranking, half-life estimate | Introduce stabilizing modifications (D-amino acids, cyclization) |
| Low membrane permeability | Membrane insertion free energy + permeability model | Permeability coefficient (Papp), insertion energy | Prioritize cell-penetrating or oral candidates |
| Immunogenicity concern | MHC binding prediction + T-cell epitope scanning | Epitope hotspot map, MHC allele risk scores | Engineer out immunogenic motifs before in vivo testing |
Synthesizing and testing every peptide analog is slow and expensive. Computational developability prediction lets you fail fast in silico, focus synthesis on the most promising candidates, and understand why a peptide is likely to fail—so you can fix it. Our integrated approach combines physics, AI, and targeted wet-lab confirmation to give you confidence in every ranking.
Yes. We support D-amino acids, N-methylation, cyclization (head-to-tail, side-chain), stapled peptides, and common non-natural residues. Parameters are generated with GAFF or analogous force fields and validated against available experimental data.
Accuracy depends on the endpoint. Aggregation and stability predictions typically show strong correlation with experiment (R > 0.7 in retrospective benchmarks). Solubility and permeability are more challenging but provide useful relative ranking within a series. We always report confidence intervals and recommend wet-lab confirmation for critical decisions.
Yes. We offer optional validation packages including ThT aggregation assays, DLS, RP-HPLC stability, and PAMPA permeability. These are performed at our partner labs or yours, and results are integrated into the final report.
We predict MHC-I and MHC-II binding, identify T-cell epitope hotspots, and flag sequence motifs associated with immunogenicity. For chronic dosing, we recommend combining computational prediction with ex vivo human PBMC assays, which we can arrange through our network.
We deliver a ranked scorecard that aggregates all endpoints with customizable weighting. You can also request Pareto frontier analysis to identify candidates that balance efficacy-related properties with developability.
Submit your project details below, and our team will respond within 24 hours.
Talk to our technical team about your project!
I Want To Talk
