Our activities ans skills cover all main aspects in computational and theoretical chemistry.

In silico Chemistry Consulting

In silico Chemistry Consulting

Having chemistry related problem or question? If you have a chemistry related problem or task, it is most likely that it can be solved or at least assisted with in silico chemistry. We will find an optimal approach fitting your particular needs and help you choose between alternatives.

In silico Hit-to-Lead Optimization

In silico Hit-to-Lead Optimization

You have your hits with promising activity, what next? Modeling and design of drug candidates for being active in lower, preferably nanomolar concentrations. CLICK below to view an example.

hit-to-lead optimization example
Computational Toxicology

Computational Toxicology

Below, please see the list of developed QSAR and ANN models, which are available for immediate prediction. All these models are also peer-reviewed and published in JRC QMRF Inventory.

Acute toxicity – inhalation Acute toxicity – oral Acute toxicity for daphnia magna Acute toxicity test – fish Acute toxicity to birds Adsorption/desorption screening Assessment of toxicokinetic behaviour: Bioaccumulation in aquatic species (BAF and BCF) Blood-brain barrier partition / human serum albumin binding Carcinogenicity (mouse) Chronic toxicity LOAEL Daphnia magna reproduction test Eye irritation (based on in vivo) Fish early-life stage toxicity test Growth inhibition study on algae In vitro gene mutation study in bacteria In vitro mammalian cell gene mutation test Mutagenicity: in vitro study in mammalian cells Persistence: abiotic degradation in air/sediment/soil/water Persistence: biotic degradation in air Prenatal developmental toxicity study Short-term toxicity to invertebrates (honey bee) Skin irritation (based on in vivo) Skin sensitisation (GPMT/LLNA) Sub-chronic toxicity study (90 day) repeated dose toxicity oral Toxicity to activated sludge
Fragment and Ligand based Molecular Design

Fragment and Ligand based Molecular Design

We analyse all possible combinations of molecular fragments and provide the structural candidates that are expected to carry desired predetermined properties.

High Level Computational Studies of Reaction Mechanisms

High Level Computational Studies of Reaction Mechanisms

Too tough? Not really! Please see our research topics.

Virtual Library Screening

Virtual Library Screening

Planning an experimental assay? Designing or expanding a dataset? Looking for similar structures to your hits? How smart can it be done? Comprehensive descriptor and fingerprint based screening of chemical libraries for selecting subsets for experimental screening and analysis of results. The subsets are chosen based on predefined requirements that are evaluated in silico (solubility, logP, toxicities, etc). Sample study: Icosagen AS (Estonia) has developed a cell assay to measure the inhibition of HPV (Human Papillomavirus) replication. Their goal was to find novel potential inhibitors as fast and low-cost as possible. After testing the start-set of ~1500 chemicals they asked UT CCG to build a novel chemical library for further measurements. We used our virtual screening technologies and built a library of 300 untested chemicals. After experimental testing 3 chemicals out of 300 were selectively active to inhibit the initial stage of HPV replication in micromolar level.

Development of computational models for prediction of properties and activities

Development of Computational Models for prediction of Properties and Biological Activities

What will be the properties/activity of such and such compounds, should we make them? Can we predict the properties beforehand?

  • We develop predictive models (QSAR, ANN, etc.) based on customer's in-house experimental data and/or public sources. We also do the collection of experimental data from literature and data bases.
  • With already existing models we can predict biological activities, environmental fate, toxicological endpoints of chemicals (and mixtures).
  • We evaluate ADME/Tox properties and medicinal effects, based on in silico models and expert judgment.
  • We perform volume calculations - structure optimization using Molecular Mechanics, QM (semi-empirical, DFT, ab initio) methods.
  • We also do dynamic modeling using Molecular Dynamics, Monte Carlo methods.
Computer Aided Drug Design

Computer Aided Drug Design

We have tools and know-how to participate in drug design projects as a partner in fields of computational design and engineering. We do scaffold based hit and lead optimizations, early screening of possible side effects and fine tuning of the small molecule candidates.

№1

Case Study 1

I have a cell assay for determination of inhibition activities and now I would like to get a meaningful chemical library for further testing and hit expansion. How to optimize my expenses and get the best possible chemical library for further testing?

Steps taken (i) at glance:

  1. Quantitative characterization of existent set of active chemicals;
  2. Preliminary modeling and trend analysis for structure vs. activity;
  3. 3D descriptor based library screen to find additional set of similar and potentially active set of chemicals
  4. Fine tuning (hit-to-lead opt.) of chemical structures: trend analysis of structural fragments vs. activity
  5. In silico design of novel potentially active molecules based on combination of fragments and extrapolation of molecular characteristics and experimental data.

Steps taken (ii) in detail:

MODULE Activity Outcome
PRELIMINARY QSAR MODEL Optimization of chemical structures, 2D and 3D descriptor calculation Descriptor/activity matrix
Selection of algorithm, training and test set, QSAR modeling, validation Statistically valid QSAR model for preliminary prediction activities for novel chemicals
IN SILICO 2D FINGERPRINT VIRTUAL SCREENING Assessment of available data. Ranking and selection of compounds by activity vs similarity analysis, assigning weights for fingerprint screening Selected active compounds for further scaffold expansion
2D fingerprint screening over a diverse set of (appr.) 400 000 structures Selection of 200 compounds for further assessment
ADVANCED 3D FINGERPRINT VIRTUAL SCREENING Analysis of the selected 200 compounds, identification of unique structural features, conversion to 3D, preliminary geometry optimization (FF level) Description of the 200 hits from 2D screen
Structure optimization and calculation of 3D descriptors, including constitutional, toplogical, quantum-chemical, etc Descriptor/structure matrix for 100 compounds
SELECTION OF HIT COMPOUNDS Detailed screening and selection of a set of top promising similar compounds based on 2D fingerprints (50) and 3D descriptors (50) Selection of 100 compounds (50+50) for in vitro assay
HIT-TO-LEAD OPTIMIZATION Analysis of activity trend over the structural fragments Potential leads
DESIGN OF NOVEL COMPOUNDS BY COMBINATION OF FRAGMENTS Definition of training set structures as combination of fragments; calculation and screening of all possible fragment combinations; selection of best potentially active compounds Set of novodesigned molecules with potential biological activity

№2

Case Study 2

Estimation of Physical Properties Problem: A company is processing fats and fatty acids at large scale. They need information of the physical properties, like melting point, density, and vapor pressure of some of their compounds at different temperatures ranging from 100°C to 300°C. Solution: based on computational chemistry: a predictive model with validated good performance is identified from the literature for predicting the vapor pressure as a function of temperature.  The model is recreated and applied to the compounds under consideration. A new predictive model is developed for the prediction of melting points based on closely similar compounds. Isothermal-isobaric molecular dynamics simulations are carried out to find the volume of the compounds at particular temperatures, providing estimations of the densities.