Advanced Modeling Suite

Molecular docking, dynamics simulations, QSAR, and pharmacophore analysis -- powered by AI

Molecular Docking Studio

Checking...

Protein-Ligand Complex

Load a Protein to Begin

Paste a PDB ID, upload a PDB file, or fetch from RCSB

Try:

Docking Parameters

Algorithm

Exhaustiveness Value: 8

Number of Modes

Grid Size (Angstrom)

Grid Center (x, y, z)

Docking Results

Load protein & ligand, then run docking

Molecular Dynamics Simulator

Checking...

Trajectory Viewer

Set Up a Molecular System

Load a protein structure, select force field parameters, and launch a molecular dynamics simulation with real-time trajectory visualization.

Quick load:

Simulation Parameters

Force Field

Temperature (K)

Pressure (bar)

Simulation Time (ns)

Time Step (fs)

Analysis Metrics

Pharmacophore Builder

Checking...

3D Pharmacophore Model

Pharmacophore Features

H-Bond Donor 3

H-Bond Acceptor 2

Hydrophobic 4

Aromatic 2

Positive Charge 1

Negative Charge 0

QSAR Model Builder

Checking...

Molecular Descriptors

Descriptor Set

Feature Selection

ML Algorithm

Cross-Validation

Training Dataset (CSV) CSV with 'smiles' and target property columns

Model Name

Target Property Column

Model Performance

R² Score 0.892

RMSE 0.234

MAE 0.156

Q² (LOO-CV) 0.845

Features Used 47 / 200

Virtual Screening Pipeline

Checking...

Select Protein Target

Choose the receptor for virtual screening

Build Compound Library

Select candidates to screen (click to toggle)

Aspirin

COX Inhibitor

MW: 180.2

Ibuprofen

NSAID

MW: 206.3

Caffeine

Stimulant

MW: 194.2

Imatinib

Kinase Inhibitor

MW: 493.6

Tamoxifen

ER Modulator

MW: 371.5

Metformin

AMPK Activator

MW: 129.2

Serotonin

Neurotransmitter

MW: 176.2

Penicillin G

Antibiotic

MW: 334.4

Lopinavir

Protease Inhibitor

MW: 628.8

Remdesivir

Antiviral

MW: 602.6

Dexamethasone

Corticosteroid

MW: 392.5

Chloroquine

Antimalarial

MW: 319.9

0 / 12 selected

Run Virtual Screening

Dock all selected compounds against the target

AI Drug Discovery Lab

Checking...

REINVENT

Transformer + Policy Gradient for de novo molecular generation

MolDQN

Deep Q-Network for molecular optimization via atom-level edits

Rainbow DQN

Double DQN + PER + Dueling + Multi-step + Distributional + Noisy Nets

Epoch 0 / 1000 0%

Status: Idle ETA: --

Training

Epochs 1000

Batch Size

Learning Rate 1e-3

RL Core

Gamma (Discount) 0.990

Epsilon Start → End

Replay Buffer

Target Update Freq

Network Architecture

Hidden Dims

Embedding Dim

Num Layers 3

Dropout 0.10

Reward Weights

QED Weight0.35

LogP Weight0.20

SA Score Weight0.25

Similarity Weight0.20

Sum: 1.00

Target Protein

Protein Target

Reference Active Compounds (SMILES) 0 reference compounds loaded

Binding Site

Box Size (A)20 A

Binding Affinity Weight0.00

Curriculum Learning

Gradually shifts from drug-likeness to binding optimization

Docking FrequencyEvery 20 epochs

Rainbow DQN

N-Step Returns3

Atom Count (C51)

Noisy Net Sigma0.50

PER Alpha0.60

PER Beta0.40

Loss & Reward

Property Tracking

Generation Quality

Best vs Target

Target Binding & Similarity

Top Generated Molecules
Best candidates discovered during training

Start training to discover molecules