Consulting & Solutions GmbH
Home Company Info Impressum Contact Email

a chemoinformatics company


AKos Consulting & Solutions GmbH is an authorized agent for CompuDrug International

New Developments


New Developments

"Topological Polar Surface Area" (TPSA)

Molecular polar surface area (PSA) is a descriptor showing the correlation with passive molecular transport through membranes, which allows prediction of the human intestinal absorption, Caco-2 monolayers permeability, and blood-brain barrier penetration.

The calculation of PSA in a classical way requires special software to generate the 3D structures that makes the prediction process is time consuming. The "Topological Polar Surface Area" (TPSA) approach (published by J Med Chem. 2000 Oct 5;43(20):3714-7., Ertl P, Rohde B, Selzer P.) allowed the high-throughput calculation of the PSA values and the quick estimation of related transport properties.

The TPSA module of Pallas frame system provides a quick high-throughput tool for PSA calculation of compound libraries, and the early identification of drug candidates with inappropriate absorption.

Web-based, network version of the EMIL software

For lead optimization purposes, CompuDrug launched the web-based version of the EMIL (Example-Mediated Innovation for Lead evolution) software. EMIL Net can be used from any terminal of your company without the need of multiple installations. The lead evolution and lead optimization databases of EMIL can be accessed via an Intranet/Internet network; all you need is to open a web-browser and login to the central EMIL server of your company. Every features of the Windows version are available in the new release, but in a renewed style satisfying the special needs of the web environment.

EMIL Net belongs to the next generation software family of CompuDrug, which provides platform-independent, integrated use of drug design tools.

MetabolExpert, HazardExpert and Rule of 5 are implemented in Pallas SDK

The Pallas Software Development Kit (SDK) has been extended with the HazardExpert, the MetabolExpert, and the Rule of 5 modules.

Pallas SDK is developed for use of system administrators and software developers. It is based on the widely-used prediction engines of Pallas, which are available in the form of shared libraries (DLL/DSO), and can be embedded into your in-house database or chemical management system. Pallas SDK 2.0 contains multithreaded DLLs and a Java Native Interface for all of the available modules, so you can also access Pallas from the Java and the C/C++ environment.

Pallas SDK allows a flexible usage of ADME/Tox prediction programs for advanced computational scientists.

MetabolExpert and RetroMEX - Colored Highlights of Fragment Reaction

The new features of MetabolExpert help the users to understand the metabolic pathway of the MetabolExpert knowledge base. The active and replacing substructures and the positive conditions are highlighted via color drawing, and the highlighting emphasizes the essence of metabolic reactions that occurred.

Newly Improved, User-Friendly Pallas:

  • Easy Access toolbars for the Pallas modules and frequently-used functions.
  • Simple prediction button and menu for large database predictions.
  • Additional modifications in Pallas applications to make them more comfortable.
  • Improved appearance of Pallas background pictures.



CompuDrug offers a range of predictive software tools to improve performance in the fields of pharmaceutical and medicinal chemistry research.

CompuDrug´s Software Solutions are providing you help in:

ADME and evaluation and prediction by Pallas
  • Rule of Five evaluation
  • Predicting metabolism pathways in humans, animals, and plants
  • Estimating toxic symptoms of organic compounds in humans and animals

In silico calculation of physico-chemical properties by Pallas

  • LogP, pKa, logD calculation

Finding hits by structural modeling, QSAR and other rational drug design approaches

  • Prodrug design by RetroMex of Pallas
  • Emil serendipity enhancer and knowledge base management

Lead Optimization

  • Emil serendipity enhancer and knowledge base management
  • Lost the in vitro activity in vivo? MexAlert of Pallas identifies first pass effect.

High Througput Screening

  • Identifying potential false positive compounds at cell based high, medium and low throughput screening assays by ToxAlert of Pallas

For the analytical lab

  • Method development and optimization for HPLC techniques by Eluex

For agrochemical research:

  • Light stability prediction of agrochemical candidates by the Agro version of Pallas Metabolexpert.

For detailed information about each product, please go to www.compudrug.com.


New Artificial Intelligence - powered accurate logP prediction: PrologP 7.0.

ANNlogP is the new prediction method used in our PrologP 7.0 software. It is based on the atomic fragment collection of Ghose and Crippen. Instead of the more common linear approaches, the new ANNlogP method uses a neural network model. Since the neural network is able to recognize the hidden and non-linear relationships between the chemical structure and the logP value, the new software provides much more accurate predicted values, and makes our PrologP 7.0 software one of the most accurate logP predictors on the market today.
The program still uses the original fragmental methods, but combines them with the new neural network algorithm, giving an optimal prediction result. The balancing between the different calculation methods was fine-tuned based on a large set of experimental logP values (approx. 13,000 compounds).

EMIL 2.3 - Expanded knowledge

The Example Mediated Innovation for Lead evolution (EMIL) software takes your screening hit and suggests chemical modifications to turn it into a bona fide lead. EMIL searches through its extensive Knowledge Base looking for similar chemistry and how it was optimized for potency and bioavailability.
The software’s Knowledge Base has been expanded remarkably. The inserted collection of data includes the results of drug discovery research of the recent years, and provides large numbers of suggested structures for the lead optimization or lead evolution process. EMIL 2.3 allows you to extend your virtual libraries with the suggested structures.

Rule of 5 - new HTS module in Pallas for Windows

Since the majority of drug candidates fail because of bioavailability problems, the early identification of poor absorption can save a huge amount of time and money. The most common screening method is based on Rule of 5.
This new Pallas module calculates the molecular weight, the logP value, and the number of the H-bond acceptor and H-bond donor groups, which characterize the drug-likeness of organic compounds. The Pallas-Rule of 5 module is developed to choose which compounds are acceptable for high-throughput screening: it is able to take an SDFile of any size as input, and presents the result in a simple tabular form that can easily be transferred to Excel.

Speed up your calculation with Pallas Cluster

Our latest product is a uniquely distributed system that is able to predict the physico-chemical parameters of huge compound databases used in high-throughput screening and combinatorial chemistry very quickly. The Pallas Cluster allows the usage of free resources of office computers for the physico-chemical property prediction of large chemical databases. The Pallas Cluster predicts the logP, logD and pKa values with the well known and widely-used Pallas prediction engines, which is embedded into a web-service based distributed system, and the power of clustered computers improves the speed of the prediction even by orders of magnitudes.

Pallas for UNIX/Linux has been upgraded

The Unix/Linux version of Pallas prediction modules has been upgraded and extended with the MetabolExpert, HazardExpert and Rule of 5 modules.

In the Unix/Linux environment the Pallas prediction modules are available as command-line software. It allows to make batch prediction for huge compound libraries, and avoids the need of the time-consuming import/export procedures. The new release accepts SMILES structure format as input, and able to export the results in SDFile format, so the predicted values can be stored together with the corresponding compound structures. This latter new feature made possible the implementation of MetabolExpert, where the predicted metabolite structures are stored in SDFile sets. An additional new function is the smiles input format.