Tel:

Email:

Services
Online Inquiry

Property Prediction of Drug-like Molecules

Inquiry

Identification of drug-like molecules is one of the major challenges in the field of drug discovery. Thus, there is a need to develop computational method that can predict property of drug-likeness molecules with precision, such as developing algorithm for screening chemical library for their drug-like properties. The ability to assess key properties of drug-like molecules using complex calculations with accuracy comparable to that of experimental laboratory assays, to facilitate optimization of drug properties, including drug potency, selectivity, and bioavailability. The Alfa Chemistry platform integrates predictive methods with various computational techniques to accelerate drug discovery. Our iterative process is designed to accelerate evaluation and optimization of chemical matter in silico ahead of synthesis and assay. The most promising compounds emerging from each round of experimental project chemistry are then further optimized through additional cycles of computation analysis.

Property prediction based on graph convolution neural networks. Figure 1. Property prediction based on graph convolution neural networks. (Wang, X.; et al. 2019)

Advantages of Property Prediction of Drug-like Molecules

  • Narrow down the number of molecules which are need to validated.
  • Assess the polymorph risk of a drug molecule to save time and effort.

Our Services

Alfa Chemistry provides easy-to-use models and computational tools for property prediction of drug-like molecules. Our rapid and high-quality services are as follow:

  • Deep learning

In order to identify drug-like molecules that can attach to disease-causing proteins and change their functionality. We apply deep leaning method to study the 3D shape of a molecule to understand how it will attach to specific surfaces of the protein.

1. Our scientists have created a deep learning model that predicts the 3D shapes of a molecule solely based on a graph in 2D of its molecular structure, in which molecules are typically represented as small graphs.

2. We have developed a message passing neural network to predict specific elements of molecular geometry. We can use it to predict the lengths of the chemical bonds between atoms and the angles of those individual bonds.

3. Our method can correctly identify the chirality by determining the 3D structure of each bond individually.

  • Ab initio prediction

Alfa Chemistry has established a cloud-native platform to predict crystal forms that are better suited for formulation with regards to physico-chemical properties, assisting in identifying the potential risk of polymorph.

1. We can rank crystal structures by energy using periodic quantum mechanics (QM) lattice calculations (e.g. plane-wave density-functional theory).

2. Our teams can also calculate the entropic contribution to crystal stability at the QM level of theory, and thereby rank the crystal structures at finite temperature.

  • Graph neural network

Solubility of drug-like molecules is related to pharmacokinetic properties such as absorption and distribution, which affects the amount of drug that is available in the body for its action. At Alfa Chemistry, we have designed a novel method based on graph neural network to predict solvation free energies.

1. We apply a message passing neural network to compute inter-atomic interaction within both solute and solvent molecules represented as molecular graphs.

2. Our experts use the features from the preceding step to calculate a solute-solvent interaction map, which captures the solute-solvent interactions along with the features such as the electronic and steric factors that govern the solubility of drug-like molecules.

3. We use the model to predict solvation free energies involving a large number of solvents with high accuracy.

  • Artificial neural network

We apply preADMET, which is one of the online resources to predict ADME, toxicity, drug likeness and molecular descriptor calculation. Our experts use Genetic Functional Approximation (GFA) to find the best descriptors set for a training set and we perform batch run neural net using descriptors in each equation obtained by GFA learning. We also evaluate the performance of trained artificial neural network and validate the final model using external dataset for testing.

  • Machine learning techniques

Our scientists have developed multiple algorithms for screening chemical library for predicting properties of drug-like properties.

1. Principal component analysis (PCA) and substructure fragment analysis are applied to perform data analysis.

2. In order to evaluate the performance of different fingerprints, we have developed various models on different sets of descriptors.

3. Alfa Chemistry supports multiple fingerprints based models such as Monte-Carlo (MC) approach, PCA based model and hybrid models.

Features of Our Molecular Simulation

  • Our models can be easy to integrate with other deep learning models.
  • Our method has potential to be applied to the area of high-throughput virtual screening, and we can utilize the model to determine small molecule structures that would interact with a specific protein.
  • Alfa Chemistry keeps refining models with additional training data so it can more effectively predict the structure of long molecules with many flexible bonds.
  • We are capable of applying various force-fields and QM energy models for property prediction of drug-like molecules.

Our property prediction of drug-like molecules services remarkably reduce the cost, promote further experiments, and accelerate the process of drug design for customers worldwide. Our personalized and all-around services will satisfy your innovative study demands. If you are interested in our services, please don't hesitate to contact us. We are glad to cooperate with you and witness your success!

Reference

  • Wang, X.; et al. Molecule Property Prediction Based on Spatial Graph Embedding. Journal of Chemical Information and Modeling. 2019, 59(9).

Alfa Chemistry

Company
  • Tel:
  • Fax:
  • Email:
Privacy Policy | Cookie Policy | Copyright © 2024 Alfa Chemistry. All rights reserved.