VIVA Multi-modal Ai-Rooted Solutions
Pep2MARS
A unified portal for cyclic peptide database development, AI design, structure generation, molecular dynamics, and conformation sampling. The entry points are organized into two tracks, Data & AI and Structure & Simulation, making it easy to move from data foundations to prediction, sampling, and analysis.
Curated structural and property data that supports both labeled and unlabeled learning workflows.
Pretraining, fine-tuning, and multimodal representations aligned with cyclic peptide property modeling.
A connected path from structure generation to MD simulation, analysis, and conformation sampling.
Data & AI
Databases, Modeling, and Developability Prediction
Start from a reliable data foundation, then connect multimodal representation learning and ADMET evaluation into a practical iteration loop for cyclic peptide property prediction.
Database Hub
Integrated Cyclic Peptide Database
A curated collection of 2D/3D structures, physicochemical properties, and amino acid libraries that provides high-quality labeled and unlabeled data for downstream model development.
Features
AI enhance capacities
Data-Driven Deep Learning Models
Pretraining and fine-tuning strategies combined with 2D-3D representation learning, including ideas such as 3D Infomax, to build AI models for cyclic peptide design.
Features
ADMET Screening
Cyclic Peptide ADMET Prediction
Rapid prediction of absorption, distribution, metabolism, excretion, and toxicity risk to support early-stage screening and prioritization of cyclic peptide candidates.
Features
Structure & Simulation
From Structure Generation to Dynamics Analysis
Connect SMILES-to-PDB structure construction, MD simulation and analysis, and conformation sampling protocol exploration into a standardized workflow for cyclic peptide structural research.
Structure Builder
Residue-Level Peptide Structure Builder
A residue-level workflow from SMILES/2D SDF to 3D PDB that generates standardized structural inputs for downstream sampling, simulation, and analysis.
Features
Molecular Dynamics
Automated MD Parameterization
Setup, execution, and batch processing for molecular dynamics simulations of cyclic peptide systems to evaluate structural stability and dynamic behavior.
Features
Trajectory Analytics
MD Analysis
Trajectory-focused analysis of conformational changes, key interactions, and statistical descriptors to derive interpretable structural dynamics insights.
Features
Conformation Sampling
Conformation Sampling
Benchmark-driven conformation generation using public datasets, with exploration of sampling protocols suitable for non-canonical amino acids.
Features
References
- 1. Li, J.; Qian, Y.* Pep2MARS: Automated Cyclic Peptide Parameterization for Molecular Dynamics and Compound Design. J. Chem. Inf. Model. 2026, 66, 6211–6217. https://doi.org/10.1021/acs.jcim.6c00340
- 2. Wen, S.; Wang, Y.; Qian, Y.* EnsembleCycPerm: Interpretable Modeling of Cyclic Peptide Permeability through Solvent-Dependent Conformational Ensembles. J. Chem. Inf. Model. 2026, ASAP. https://doi.org/10.1021/acs.jcim.6c01213