Use ESPsim Online
Commercially Available ESPsim No-Code Web Server
ESPsim: 3D Shape and Electrostatic Similarity for De Novo Drug Design
Maximize your multiparameter optimization with ESPsim (ESP-Sim), an open-source tool now fully integrated into Tamarind Bio for advanced comparison of electrostatic potentials and molecular shapes. ESPsim overcomes the traditional bottlenecks of 3D deep generative methods, allowing drug hunters to seamlessly identify bioisosteres and guide fragment-based molecular generation without needing a prior accurate description of bioactive conformations.
Why ESPsim?
Multiparameter optimization is the heart of automated drug design. While traditional generative methods rely heavily on text-based 2D SMILES representations, small modifications in text strings can catastrophically alter a molecule's actual 3D structure.
Because molecular recognition is primarily driven by three-dimensional shape and electrostatic properties, ESPsim evaluates similarities where they matter most. By computing overlap integrals of Coulomb potentials, ESPsim provides a highly sensitive, physics-based metric to bias deep generative models toward realistic lead analogs and highly potent "sweet spot" molecules.
Key Features & Innovations:
Conformation-Independent 3D Fragment Comparison: Breaks down single bonds extending from ring atoms, aligning pairs of fragments using an automated ETKDG conformer generation system coupled with a rigid internal anchor template (e.g., a hexazine ring). This entirely bypasses the challenging step of generating bioactive conformations for complete target molecules.
High-Quality Charge Calculation Integration: Supports Gasteiger, MMFF94, predicted machine learning (ML) partial charges, AM1-BCC charges, or high-fidelity quantum mechanical Restrained Electrostatic Potential (RESP) charges computed via Psi4 (such as B3LYP/6-31G).
Flexible Scoring Metrics: Provides electrostatic potential similarity evaluated via Carbo or Tanimoto metrics, alongside volumetric shape Tanimoto scores and an integrated ESP-TanimotoCombo score.
Scaffold-Hopping Power: Explicitly rewards the discovery of novel central heterocyclic rings that match known leads in shape and electrostatics while maintaining intellectual property (IP) diversity.
Validated Performance Benchmarks
ESPsim has been rigorously evaluated across standard protein-docking databases, rescoring tasks, and scaffold-hopping comparative benchmarks:
Partial Charge Robustness: When compared to quantum mechanical RESP ground truth (MP2/cc-pVTZ) across ~3,000 neutral molecules, AM1-BCC charges reproduced QM electrostatic potentials best, followed by deep learning (ML), MMFF, and Gasteiger charges.
Industry Alignment: ESPsim scores demonstrate a high correlation with state-of-the-art closed-source tools like EON, yielding a Spearman correlation coefficient of ~0.8 across molecular fragments.
Target Enrichment: Benchmarking against the 102 DUD-E targets and dopamine D4 receptor rescoring functions confirmed that ESPsim electrostatic and shape similarities excel at identifying true active ligands against key protein targets.
Real-World CDK2 Case Study: In a retrospective study to identify bioisosteres for CDK2 kinase inhibitors, ESPsim successfully ranked the equipotent imidazopyridine and pyrazolopyridine fragment pair as the top matching replacement, significantly outperforming conventional 2D fingerprints (such as Morgan, MACCS keys, and topological fingerprints).
What is Tamarind Bio?
Tamarind Bio is a pioneering no-code bioinformatics platform built to democratize access to powerful computational tools for life scientists and researchers. Recognizing that cutting-edge machine learning and quantum chemistry packages can be highly complex to deploy, configure, and scale, Tamarind Bio abstracts away the difficulties of high-performance computing, software dependencies, and command-line interfaces.
By managing backend orchestration, parallel computing (via libraries like Dask), and GPU scaling on a secure, cloud-hosted environment, Tamarind Bio enables drug discovery teams to focus purely on scientific intuition and rapid molecular iteration.
How to Use ESPsim on Tamarind Bio
Leverage the power of ESPsim in your next drug discovery campaign with a streamlined, zero-setup workflow:
Access the Platform: Log in to your secure account on the Tamarind Bio website.
Select ESPsim: Locate and select the ESPsim model tool from the dashboard menu.
Upload Your Molecular Inputs: Input your target lead molecules or fragment pairs in a standard format (such as SMILES or SDF).
Configure Your Parameters:
Select your preferred partial charge method (e.g., Gasteiger, MMFF94, AM1-BCC, or quantum mechanical Psi4 RESP charges).
Choose your alignment parameters, anchor template groups, and scoring metrics (Tanimoto, Carbo, or TanimotoCombo).
Run the Analysis: Click submit to execute the simulation. Tamarind Bio automatically handles the parallelized cloud architecture to process complex calculations.
Analyze & Export: View interactive 3D electrostatic contours, review calculated similarity matrices, and download high-fidelity reports or structure files for your downstream synthesis and testing.