Pharma Algorithms

June 9, 2008

Pharma Algorithms releases Tox Boxes version 2.9 Toronto, Canada Pharma Algorithms Inc. announces the release of Tox Boxes v 2.9. This release contains two new modules allowing pharmaceutical researchers to address possible drug safety issues related to hERG Inhibition and CYP3A4 Inhibition. In addition Genotoxicity module has been updated to include Genotoxicity Hazards knowledge-based expert system. ‘Trainable Models’ section features trainable versions of the newly introduced modules (Trainable hERG Inhibition and Trainable CYP3A4 Inhibition) as well as the revised and improved ‘Trainable Models’ available from earlier versions of Tox Boxes. Updates have also been introduced in the suite of LogP, pKa, LogD and solubility prediction modules available for initial assessment of ADME profile. Along with the increased prediction accuracy, some of them contain newly introduced features such as additional solubility threshold in Qualitative Solubility in buffer module. Request the trial version of TOX Boxes by filling out the short registration form.

December 20, 2007

Pharma Algorithms releases Tox Boxes version 2.5 Toronto, Canada Pharma Algorithms Inc. announces the release of Tox Boxes v 2.5. This new version introduces an Aquatic Toxicity module providing capabilities to estimate compound concentrations lethal to the Fathead minnow (Pimephales promelas) and Water flea (Daphnia magna) species. In addition this software package upgrade contains implementation of the ‘Trainable Model’ concept. This concept, utilizing a novel similarity based methodology, allows the user to: - Quantitatively assess the quality of the predictions by means of the Reliability Index estimation. - Instantly retrain the model with the help of any user-defined, proprietary or ‘in-house’, database of experimental values for the property of interest. These ‘Trainable’ tools, allowing for expansion of the Model Applicability Domain, open new possibilities for the use of Tox Boxes predictive algorithms in the chemical industry. The software also includes calculation of LogP, pKa, LogD and solubility for initial assessment of ADME profile. For this release, the following modules provide both Reliability Index estimations and the ability to retrain those using user-defined data: - Genotoxicity (Ames Test) - Quantitative Solubility in pure water (LogSw) - Quantitative Solubility in buffer (LogS) - Qualitative Solubility in buffer - Octanol-water or buffer partition coefficients (LogP & LogD) Additionally, predicted Reliability Index values are available in Acute Toxicity/LD50 and Aquatic Toxicity modules. Request the trial version of TOX Boxes by filling out the short registration form.

November 22, 2007

Pharma Algorithms releases ADME Boxes version 4.0 Toronto, Canada Pharma Algorithms Inc. announces the release of ADME Boxes v 4.0. In this release three new modules have been added to address the needs of pharmaceutical researchers to explore compound permeability across Blood-Brain Barrier or Caco-2 monolayer and Qualitative Solubility in buffer. Along with these entirely new modules this major upgrade of the software package marks the introduction of important innovations in the methods behind ADME Boxes predictive models. The development of the ‘Trainable Model’ concept utilizing a novel similarity based methodology allows the user to: - Quantitatively assess the quality of the predictions by means of the Reliability Index estimation. - Instantly retrain the model with the help of any user-defined, proprietary or ‘in-house’, database of experimental values for the property of interest. These ‘Trainable’ tools allowing expansion of the Model Applicability Domain open new possibilities for the use of ADME Boxes predictive algorithms in the chemical industry. Current modules upgraded to provide Trainability and Reliability Index - P-gp Substrate/Inhibitor Specificity - Plasma Protein Binding (LogKa & %PPB) - Quantitative Solubility in pure water (LogSw) - Quantitative Solubility in buffer (LogS) - Qualitative Solubility in buffer - Octanol-water or buffer partition coefficients (LogP & LogD) Request the trial version of ADME Boxes by filling out the short registration form.

June 26, 2007

Pharma Algorithms releases ToxBoxes version 2.0 Toronto, Canada New developments in this major Tox Boxes upgrade provide researchers with significantly enhanced and expanded Acute Toxicity prediction capabilities and two completely new modules for the assessment of a compound’s potential to cause acute eye or skin irritation and disruption of endocrine system. The software also includes calculation of LogP, pK_a, LogD and solubility for initial assessment of ADME profile. New functionality is provided through knowledge based “expert-like” approaches, identifying and utilizing substructures known to influence the property of interest, e.g. toxicophores in Acute Toxicity module. An innovative methodology has been implemented to provide the predictions with a reliability estimation based on experimental values for the similar compounds in the training set. This Reliability Index is an important milestone toward a main REACH requirement – the possibility to assess the applicability domain of in-silico models. Acute Toxicity module provides evaluation of compound toxicity in three different and independent ways: - LD₅₀ prediction with estimation of the prediction reliability (Reliability Index); - Probabilistic model for classification of compounds into OECD derived toxicity categories based on the thresholds of numeric criteria expressed as LD₅₀ (oral administration in rats); - Toxicity Alerts – Identification of hazardous fragments that may be responsible for the high acute toxicity of compounds in rodents with toxicity distribution plots supporting the significance of the identified hazardous fragments on the acute toxicity. Irritation module combines statistical and rule-based approaches to evaluate the probability for a compound to cause moderate to severe eye and skin irritation in a standard Draize test. Also provides differentiation of the irritational properties of organic electrolytes and their salts. Endocrine Disruption module contains a substructure-based binomial PLS model that classifies compounds according to their affinity to the estrogen receptor α. Ames Test module predicts mutagenicity potential of the analyzed molecule in the Ames test – one of the most popular tests for assessing mutagenic properties of the compounds in drug and chemical industry. The module provides identification of the parts of the molecule associated with mutagenic activity. Tox Boxes predictive models have been implemented as automated software applications requiring only chemical structure input with a straightforward graphical user interface designed to meet the needs of toxicologists and medicinal chemists who require interactive functionality as well as computational chemists who need to perform virtual screening and high-throughput property filtering of virtual libraries. ToxBoxes [more info]

October 31, 2006

Pharma Algorithms releases ADME Boxes version 3.5 Toronto, Canada Pharma Algorithms Inc. announces the release of ADME Boxes v 3.5. In this release, three new modules have been added to address the need of pharmaceutical researchers to explore compound’s pharmacokinetic profile and recommended dosing. The new pK_a module now calculates ionization constants for all stages and simulates distribution plots for protonation and group states. All these modules can be accessed interactively as well as in a high-throughput mode. ADME Boxes [more info]

February 1, 2006

Pharma Algorithms releases an updated DMSO solubility predictor version 2.0 Toronto, Canada Pharma Algorihtms Inc. announces the release of DMSO Box 2.0. Prediction of DMSO solubility plays a vital role in sample management for drug discovery. The distribution of samples as DMSO solutions is preferable to distribution of solids or powders. Compounds that are insoluble in DMSO present considerable difficulties to standard laboratory handling and HTS protocols and are costly to develop. Plating of these compounds is severely hampered and may even lead to erroneous measurements. The DMSO predictor uses a Specs data set of more than 20,000 compounds with a solubility threshold of 20 mM. This set contains diverse drug-like structures and was obtained by plating out more than 135,000 compounds over a period of 10 years. The updated predictive algorithm utilizes Pharma Algorithms’ molecular property prediction technology based on dynamically defined molecular fragmentation. It predicts the likelihood of a compound’s DMSO solubility exceeding a threshold of 20 mM, thus providing a quantitative criterion that can be used to rank or filter compound candidates prior to the actual screening experiment. The algorithm has been implemented as a fully automated software application that runs both interactively and in a high-throughput batch mode. The software has a straightforward graphical user interface designed to meet the needs of scientists who participate in screening projects and need to perform virtual screening and high-throughput property filtering of virtual libraries DMSO Solubility [more info]

January 25, 2006

Pharma Algorithms releases ADME Boxes version 3.0 Toronto, Canada Pharma Algorithms Inc. announces the release of ADME Boxes v 3.0. In this release, three new modules have been added to address the need of pharmaceutical researchers to model compound absorption and intestinal permeability, protein-bound fraction in blood plasma and volume of distribution. The new modules can be accessed interactively as well as in a high-throughput mode. The new functionality builds on Pharma Algorithms’ molecular property prediction technology based on dynamically defined molecular fragmentation, adding mechanistic modeling of absorption processes such as different routes of permeability and different rates for multiple ionic forms of a compound. The predictive models have been implemented as automated software applications with a straightforward graphical user interface designed to meet the needs of medicinal chemists who require interactive functionality as well as computational chemists who need to perform virtual screening and high-throughput property filtering of virtual libraries. Absorption module contains a mechanistic predictive model of human intestinal permeability. The predictions take into account the transcellular and paracellular routes of permeability and different rates for different ionized forms of a compound. The module can be used to predict the following absorption-related properties: 1. Maximum achievable human intestinal absorption (when solubility/dissolution is not a limiting factor); 2. Jejunum permeability; 3. Caco-2 permeability; 4. Absorption rate (k_a). In the Absorption module, LogP and pKa values that are used in the predictive algorithm can automatically be calculated by the program from compound structure or entered manually if their reliable experimental values are available. Researchers can also explore the effects of deliberate changes to compound analogues’ lipophilicity and ionization constants on the intestinal absorption profile. By entering modified LogP or pKa values in the graphical interface of the module, researchers can model the changes in compound absorption and estimate the magnitude of the effects. Protein Binding module predicts plasma protein bound fraction and the equilibrium binding constant to blood serum albumin of a compound in blood. The protein binding properties are predicted from automatically calculated physicochemical properties such as lipophilicity, ionization constants, and hydrogen bonding capacity. Volume of Distribution module contains a predictive model which generates a quantitative estimate of the apparent volume of distribution of a compound. Physicochemical parameters, charge state, lipophilicity and hydrogen bonding capacity are automatically calculated and used as inputs to the predictive model of the volume of distribution. The Absolv module algorithm has been updated to increase the accuracy of predictions. The user interface has also been enhanced: contributions of each atom to the currently selected Abraham parameter are color-mapped onto the structure, with intensity of the color indicating the degree of contribution of each atom or substructure to the selected parameter. ADME Boxes [more info]

January 16, 2006

Pharma Algorithms releases Algorithm Builder version 1.8 Toronto, Canada Pharma Algorithms announces the release of Algorithm Builder, version 1.8. This version adds: - A new substructure matching method which allows Algorithm Builder users to perform simultaneous match counts of multiple molecular fragments and predefined fragment sets. It is now possible to create and experiment with custom molecular fingerprints of any length or composition. The new method, Fixed-fragment Fragmentation, generates counts of occurrences of predefined structural fragments or generalized templates in all compounds in a database. It can be run with static as well as dynamically generated fragment sets in conjunction with another fragmentation method, taking the list of fragments generated by that method as its input. - Advanced substructure matching and molecular fingerprinting with a command-line utility without launching Algorithm Builder. Fixed-fragment fragmentation settings can first be exported from Algorithm Builder into a self-contained file which is then used by the command-line Fragmentation Launcher to generate a text table of fragment match counts in structures in a virtual compound library independently of Algorithm Builder. Algorithm Builder [more info]

October 24, 2005

Pharma Algorithms releases ADME Boxes version 2.5 Toronto, Canada Pharma Algorithms Inc. announces the launch of ADME Boxes 2.5. This release includes: - Unlimited batch as an optional module - The ability to create and manipulate structure dictionaries to help organize virtual compound libraries - LogP module with substructure highlighting and display of close analogues to assist with property space visualization - Flexible and configurable navigation between modules

June 15, 2005

Pharma Algorithms releases Algorithm Builder version 1.7 Toronto, Canada Pharma Algorithms announces the release of Algorithm Builder, version 1.7. This version brings new and enhanced statistical inference methods combined with simplified production-quality model deployment to project teams. Some of the new features that were thoroughly tested and are now available with this version: - Binomial PLS - Auto-determination of the optimum number of principal components by iterative sampling (bootstrapping) - Identification of robust variables by bootstrapping in combination with PLS or binomial PLS - More parameters available to estimate model robustness - Simplified molecular structure and compound ID handling

May 23, 2005

Pharma Algorithms releases a workgroup version of its ADME/Tox software suite, ADME/Tox WEB Toronto, Canada Pharma Algorithms announces ADME/Tox WEB, a workgroup version of ADME and Tox Boxes. ADME/Tox WEB has been specifically designed to fit into departmental and corporate collaborative networks. Users can access the prediction tools at any time and from any computer on the organization's intranet without additional installation. ADME/Tox WEB offers its owners powerful interface customization capabilities and direct programmatic access to its modules. The WEB suite predicts all of the following: - ADME: overall oral bioavailability, pKa, logD at any pH, P-gp substrate and inhibitor specificity, solubility in pure water and in buffer, Abraham solvation parameters, active transport properties, intestinal absorption, physicochemical properties (TPSA, logP, number of rotatable bonds, number of h-bond donors and acceptors); - Toxicity: acute toxicity (mouse and rat), genotoxicity, health effects (blood, cardiovascular, gastrointestinal, kidney, liver, lungs). ADME/Tox WEB [more info] Request a demo account of ADME/Tox WEB by filling out the short registration form

February 17, 2005

Pharma Algorithms releases Tox Boxes version 1.0 Toronto, Canada Pharma Algorithms Inc. announces the release of Tox Boxes v 1.0. The software provides predictions for three basic toxicity endpoints: acute toxicity, genotoxicity and organ-specific health effects. Predictions are supplied with the 95% confidence intervals or probabilities, providing an indication on the reliability of each prediction. Toxicophores are identified and displayed, giving insight as to which parts of the molecule are responsible for the toxic effect. The software also includes calculation of LogP, pKa, LogD and solubility for initial assessment of ADME profile. Tox Boxes [more info]

July 1, 2004

ADME Boxes v 2.2 - New software predictor available in Europe from Sirius Toronto, Canada Sirius Analytical Instruments Ltd (Forest Row, UK) has been appointed European distributor of ADME Boxes, a software product from Canadian-based software developer, Pharma Algorithms. ADME Boxes is a modular software product for molecular discovery providing innovative prediction tools and impressive database for oral bioavailability, P-gp specificity, solubility, solute properties, and ionisation. One of the modules is the new Absolv, the improved and enhanced version of Sirius Absolv, designed to calculate molecular properties using the principles introduced by Dr. Mike Abraham of University College London.

Sirius Analytical Instruments Ltd is the Sussex, UK-based developer and manufacturer of a range of tools for physicochemical property analysis, providing analytical instruments and services for drug characterisation. For more information about Sirius please contact Sirius@sirius-analytical.com or call +44 (0)1342 820720.

June 1, 2004

Pharma Algorithms releases ADME Boxes version 2.2 Toronto, Canada Pharma Algorithms Inc. announces the release of ADME Boxes v 2.2. In this release the software has been updated to include the following new modules and features: Absolv Module - the product of cooperative development between Pharma Algorithms and Prof. M.H. Abraham, Absolv has been redeveloped from the database up. - Prediction of Abraham solvation parameters modelled from the Abraham Database: H-Bonding acidity, H-Bonding basicity, partition coefficient between gaseous phase and hexadecane, polarity/polarizability, excessive molar refraction, McGowan volume - Predictions are provided with 95% confidence intervals - Enables calculation of various solvation-associated properties from Abraham type equations - Absolv Database module provides a fully searchable database of more than 5000 compounds Batch Calculation Module - Allows batch calculation of all the properties available in the ADME Boxes modules - Small batch allows processing of files containing 50 structures - File designation through simple dialog box

January 15, 2004

Pharma Algorithms releases ADME Boxes version 2.0 Toronto, Canada Pharma Algorithms announces the release of ADME Boxes v2.0. In this release the software has been updated to include the following new features and substantial improvements: - Human oral bioavailability (%F): prediction cut-offs at 30% and 70%, expert system of alerts - Solubility in water and in buffer: doubled training set, pH-dependences added - Stability: susceptibility to acid hydrolysis in stomach - Human intestinal permeability: prediction cut-offs at 30% and 70% - 1st Pass Metabolism: probability of metabolic transformations in liver and intestine - Active transport: PepT1, ASBT, and other transporters - P-gp transport and inhibition: knowledge-based and probabilistic predictions - Ionization: improved pKa predictions, log D-pH dependences - Searchable databases: %F, HIA, P-gp transport and inhibition, solubility

June 16, 2003

Pharma Algorithms releases ADME Boxes version 1.6 Toronto, Canada Pharma Algorithms announces the release of ADME Boxes v1.6. The software for ADME/Tox and Physical Properties calculation features an informative, unified interface with component prediction Modules. Current Modules include: Acute Toxicity(LD50), Aqueous Solubility, Human Intestinal Permeation, Ionization, P-gp Substrate Specificity. In this release the software has been updated to include the following new features: - Toxicity module: a new 'Health Effects' feature highlights parts of a molecule contributing to the toxic effect. - P-gp module: updated and new algorithms based on an expanded dataset. - Ionization module: LogD calculation has been added. - Reference Compound Dictionary - search or enter a chemical name to calculate properties for reference structure.

May 19, 2003

Pharma Algorithms and Specs collaborate to introduce novel DMSO solubility predictor on Algorithm Builder (AB) platform AAPS Workshop on Pharmaceutical Profiling, Whippany, NJ, USA Specs, the leading provider of chemistry solutions for drug discovery has combined its vast qualified database of drug-like compounds with Pharma Algorithms computational algorithm development technology and expertise to produce the first of its kind compound solubility predictor for DMSO. Click here to go to the full pdf version (84 KB) of the press release DMSO Solubility [more info]

May 7, 2003

Pharma Algorithms releases Algorithm Builder (AB) version 1.6 Toronto, Canada Pharma Algorithms announces the release of Algorithm Builder (AB) version 1.6. The new release features - Parameters - giving advanced flexibility to develop algorithms utilizing multiple resources. Now users can access any number of external fields - including experimental data and third party descriptors - when building in-house algorithms. Additional new features provide: Scaffold-based Fragmentation - a novel fragmentation method that splits structures in a data set into scaffold fragments. Similarity Search - enables the search for similar structures in a data set using selected fingerprint.

November 2002

Pharma Algorithms releases ADME Boxes version 1.0 Toronto, Canada Pharma Algorithms announces the release of Box versions of its ADME (& Tox) filters. The ADME Boxes allow users to import or paste standard 2-D structure files or SMILES strings and instantly obtain predicted values, structural analogs, reference data, and calculation reliability estimates for single compounds. Box versions are available for: Human Intestinal Permeation, Aqueous Solubility, Acute Toxicity and Ionization. These can be purchased individually or as a complete package. All boxes also include calculated LogP, No. of Rotatable Bonds, TPSA, H-Bond donors and acceptors as well as Molecular Weight.

May 2002

WILEY-VCH Journal, Quanititative Structure-Activity Relationships publishes an article by Pharma Algorithms' scientists. "Fragmental Methods in the Design of New Compounds. Applications of the Algorithm Builder" takes an in-depth look at the applicability of fragmental methods in algorithm development and provides a detailed analysis of the scientific basis underlying the software. The article was written by Alanas Petrauskas, CEO of Pharma Algorithms, Pranas Japertas, Head of Development and Remigijus Didziapetris, ADME/Tox Group Leader. Click here to go to the abstract. The full-text version is available to Wiley InterScience subscribers.

January 2002

Pharma Algorithms announces release of Algorithm Builder version blue

December 2001

Pharma Algorithms presents poster 'Analysis of ADME/Tox Data by AAB' at SCI meeting 'ADME: Perspectives in HT and in Silico approaches' London, UK

October 2001

Pharma Algorithms participates in AAPS 2001, Denver with new distributor pION,Inc of Boston, Mass Pharma Algorithms announces pre-release version of Algorithm Builder to the pharmaceutical community