Alfa Chemistry provides access to molecular property prediction services for worldwide customers. Our platform with state-of-the-art methods and highly expert computational chemists is available to promote the development of chemical and pharmaceutical fields. We guarantee the best services for you.
Molecular property prediction, a powerful way to evaluate, select, and generate molecules, is significant for drug discovery and substance discovery. Accurate prediction shortens the time and saves cost to accelerate the overall process of drug discovery. Thus, the key is to find suitable and effective computational methods. Traditional methods based on extracting fingerprints or hand-engineered features in combination with machine learning algorithms represent biases from experts to some extent. To improve this, different types of machine learning algorithms, especially deep-learning algorithms which accelerate computational power and enhances the availability of large data sets, suffice for learning representations of a specific task in an automated way. Recent studies show that Multilevel Graph Convolutional neural Network, which considers much rule-based information, has a good performance in analyzing the complex relations and structures among a large collection of molecular compounds. The advanced algorithms for molecular property prediction are useful and time-saving, particularly predicting large quantities of molecules, which speeds up drug discovery.
Figure 1. The architecture of MGCN. (Lu, C.; et al. 2019)
Our Molecular Property Prediction Services Include,
Density of charge distribution can be measured by charge distribution prediction which is an important tool to study the physical and chemical properties of many molecules. Our scientists have designed various complex algorithms including first- and last-passage, first-principles calculation and last-passage Monte Carlo algorithms to calculate charge distribution of atoms, molecules and materials, predict volume charge density, surface charge density and linear charge density.
As an important physical quantity, dipole moment is used to characterize the polarity of a molecule. At Alfa Chemistry, multiple molecular function including Hartree-Fock (HF), Moller-Plesset's second-order perturbation (MP2) and fourth-order perturbation (MP4 (SDQ)) are available to investigate the interaction between polar molecules. Our teams provide various simulation technologies and calculation methods to accurately predict the molecular dipole moment and polarizability, assist in judging physical-chemical properties about molecular polarity.
Electron affinity reflects the ease with which an atom or molecule gains an electron. Ionization potential refers to the minimum amount of energy required to remove the most loosely bound electron of an isolated neutral gaseous atom or molecule. We can provide powerful parameters for reactions and high-performance materials by analyzing molecular electron affinity and ionization potential. Moreover, our experts offer Fock-space coupled cluster (FSCC) calculation, Quantum electrodynamics coupled-cluster (QED-CC) calculation and Multi-configuration Dirac-Hartree-Fock (MCDHF) calculation to compute electron affinity and ionization potential.
As a fundamental property of atoms, molecules, and condensed phases of matter, electron density can measure the probability of an electron being present at an infinitesimal element of space surrounding any given point. We use the obtained electron density to determine the properties such as charges, dipoles and electrostatic interaction energies. A diversity of machine learning technologies which contain different functions are applied to calculate electron density of atoms, molecules, plasma, crystal (surface), etc. And other simulation and calculation methods are also ready to support the prediction of electron density.
Localization of electrons describes the extent to which its motion is trapped in a particular spatial range. We use Hartree-Fock theory to provide electron localization function calculation of atoms, molecules, and solid systems. In addition, based on ELF, we can provide topological analysis, dynamic changes analysis of chemical bonds and electron dynamics, and study atomic shell structures as well as the position of the lone pair.
Electrostatic potential is a scalar field that describes an electrostatic field. It refers to the needed work energy to move a unit of charge from a reference point to the specific point in an electric field. Electrostatic potential plays an essential role in understanding the interaction between molecules, the reaction sites and molecular recognition. Our teams use Ab initio methods, Gaussian calculation and various simulation technologies to explain and predict electrostatic interactions between molecules, calculate the electrostatic potential for molecular simulation, predict the condensed phase properties of the system.
Energy band structure describes the law of electron movement in solid-state physics. It is a significant attribute to evaluate materials and plays a crucial role in solid electron research. Energy band structure prediction for semiconductors provides crucial information about the band gap and band-edge energies. Alfa Chemistry supports multiple calculation methods and modeling techniques to provide high-quality energy band structure prediction services.
We use molecular vibrational frequency analysis to characterize molecular structures, chemical composition, chain orientation and describe the microstructure of materials. Our experts focus on studying various molecular vibration such as symmetric and asymmetric stretching, bending, scissoring and rocking, and out-of-plane. Electron energy spectroscopy is commonly applied to investigate various electronic energy spectra and monitor microstructure of materials.
Molecular orbitals describe the state of electrons in polyatomic molecules. We study the native of molecules by molecular orbitals visualization, molecular orbitals calculation, prediction of property affected by molecular orbitals. At Alfa Chemistry, diverse theoretical methods are available to provide molecular orbital analysis, calculate the potential energy surfaces of chemical reaction, and explain chemical reactivity.
Molecular structure presents the three-dimensional arrangement of the constituent atoms. And it shows the position of atoms in space and the type of bonds, such as bond lengths, bond angles, dihedral angles, torsional angles, and other geometrical parameters. Molecular structure affects multiple molecular properties and the analysis of molecular structure helps to predict the molecular interactions and understand the biomolecular structure-function relationship. Our experienced scientists provide chemical isomers calculation and molecular volume calculation to predict the boiling point of molecules and flash point of the mixture.
Nonlinear optical materials have been widely applied in various fields including laser technology, spectroscopy, and material structure analysis. Our teams provide nonlinear optical properties analysis services using multiple methods such as first-principles calculation, material high-throughput screening, and material properties prediction. We are committed to shortening the research cycle of material preparation and improving the successful possibility of desirable materials through predicting properties of nonlinear optical materials, analyzing interactions between light and matter.
Researchers use many thermochemical properties to study the thermal effect in physical and chemical transformations. We are capable of performing Ab initio methods, coupled cluster single-double and perturbative triple (CCSD(T)) methods, density functional theory methods, first-principles phonon calculations, hartree-Fock methods, and so on to provide accurate predictions of heat capacity, enthalpy, entropy, Gibbs free energy, heat of combustion, and heat of formation.
Carrier mobility is one of the most important parameters of any semiconductor material and determines its applicability in various electronic devices including FETs. Alfa Chemistry performs first-principles calculations and provides computing services to customers to calculate carrier mobility.
Orbitals are invaluable in providing a bonding model within a molecule or between a molecule and a surface. Molecular orbitals are where most computational chemistry methods now begin. Alfa Chemistry provides molecular orbital energy prediction services to its clients to solve their various application problems.
The geometry or conformation of a molecule is often referred to as the three-dimensional coordinates of its atoms. Conformation generation is the process of predicting the potential effective coordinates of a molecule, which is crucial for figuring out the chemical and physical properties of a molecule. Alfa Chemistry provides molecular geometry conformation prediction services to its clients, using a variety of computational methods to generate conformations to aid their applications.
Molecular dynamics (MD) is a computer simulation method for analyzing the physical motion of atoms and molecules. The technique involves generating atomic trajectories of a system using numerical integration of Newton's equations of motion to obtain specific interatomic potentials and boundary conditions defined by initial conditions. To advance the study of mechanics, energy, and complex biological systems using cutting-edge MD simulation techniques, Alfa Chemistry offers MD simulation services.
Our Features
- Density functional theory is a quantum mechanical method for studying the electronic structure of multi-electron systems. We use it to investigate the properties of molecules and condensed matter, and provide accurate prediction of molecular properties.
- Our molecular property prediction platforms supports the application of Multilevel Graph Convolutional neural Network (MGCN). We represent each molecule as a graph to preserve its internal structure and use MGCN to directly extracts features from the conformation and spatial information followed by the multilevel interactions.
- Some simple atom and bond attributes are applied to build atom-specific feature vectors. And various molecules are treated as undirected graphs with attributed nodes and edges to identify important features of molecular property prediction.
Alfa Chemistry provides fast, specialized, high-quality services of molecular property prediction at competitive prices for global customers. Personalized and customized service of molecular property prediction satisfies innovative scientific study demands in drug discovery and materials. Our clients have direct access to our staff and prompt feedback to their inquiries. If you are interested in our services, please contact us for more details.
References
- Lu, C.; et al. Molecular property prediction: a multilevel quantum interactions modeling perspective. Proceedings of the AAAI Conference on Artificial Intelligence 2019, 33(1): 1052-1060.
- Wieder. O.; et al. A compact review of molecular property prediction with graph neural networks. Drug Discov. Today: Technol. 2020, https://doi.org/10.1016/j.ddtec.2020.11.009.