ACD/pK_a DB

What's New

Version 10.0 to 11.0

Among the key new features of ACD/pK_a DB presented below, are:

• Enhancements to the pK_a prediction model
• A new non-aqueous solvent reference database
• ACD/Quick Report

Enhancements to pK_a Prediction Model

Version 11 of the pK_a model boasts the addition of >2,000 pharmaceutical lead compounds to the internal database. Analysis of the new subset of compounds along with a portion of the existing dataset generated over 600 new and revised Hammett equations. Improvements to the pK_a model will have a significant effect on our logD and solubility modules since pK_a is integral to these modules.

Non-aqueous Solvent Reference Database

The new reference database contains experimental data for more than 2000 molecules in solvents such as Dimethyl sulphoxide, Acetonitrile, Methanol, Acetone, and Pyridine. Records include original literature sources and experimental parameters.

ACD/Quick Report

A new PhysChem interface has been developed for version 11 that offers a snapshot of the molecular property profile of individual molecules through customizable, color-coded histograms. Quick Report includes all descriptors available to the user from any ACD/Labs commercial PhysChem product.

Changes to training:

• The ability to use multiple accuracy extender training files enabling users to combine and extend training across projects, and extend the applicability of training in different modules
• Extended application of pK_a training files to logD and solubility algorithms
• Support for import of SMILES into pK_a and PhysChem Accuracy Extender

Other improvements to ACD/pK_a DB include:

• Push data directly from Table View to Microsoft Excel at the click of a button
• Addition of 8 descriptors to the PhysChem History which depict the molecular composition of molecules
• Automatically remove explicit hydrogen atoms from structures on import of a file
• Search specific text in both user data and/or notes in database records

Download a PDF copy of the expanded details on What's New with ACD/pK_a DB, or contact your Account Manager or Distributor.

Version 9.0 to 10.0

Among the key features of ACD/pK_a DB presented below are:

• Enhancement of the fragment-breaking algorithm
• Atom Numbering Consistency between User files and the History Window

Enhancement of the algorithm
Version 10.0 development has addressed limitations of the fragment-breaking algorithm to allow pK_a to be calculated for larger/more complex molecules, such as fused ring systems. Fragmentation has now been focused around reaction centers to minimize the number of fragments generated for a given system.

Atom Numbering Consistency
Manually assigned atom numbering in ACD/ChemSketch, or numbering schemes originating from SDfiles or molfiles, will be retained in the PhysChem History and PhysChem Databases after calculation of pK_a. Calculated values will now be assigned to atoms and reaction centers by means of your own numbering system. This consistency will also extend to data exported for further evaluation or manipulation.

Other improvements to ACD/pK_a DB include:

More detailed reporting of pK_a calculation to allow facile interpretation of results

• Enhancements to fragment affiliation with pK_a calculation details—improvements have been made to the affiliation (by highlighting) of the calculation details with heteroatoms in certain aromatic systems so that visual association of incremental ΔpK_a values, sigma values, and other calculation details is more complete for a wider range of compounds.
• Reporting of sigma value origin—to give expert users more information about the derivation of ΔpK_a values, the origin of substituent sigma values are reported as experimental or calculated in the Details window, with references to experimental values provided for further investigation.

Ease of use enhancements to pK_a System Training

Enhancements made to pK_a System Training in version 10.0 will benefit users that optimize our pK_a model to increase the accuracy of prediction for novel and proprietary compounds by incorporation of experimental data.

• Flexibility in the application of User databases—you can now choose to simultaneously use fragments from both the user database (pkauser.dat) and the internal calculation database for calculation of pK_a values. The default option to apply fragment increments from the User database first is also available.
• Resize dialog boxes in pK_a system training—the ability to resize the Add pK_a Value and Edit pK_a Value dialog boxes in the Database window will particularly impact users working with very large molecules by giving you a clearer view of the structure and allowing easier identification of dissociating atoms and reaction centers.

Version 8.0 to 9.0

Among the key new features of ACD/pK_a DB presented below, are:

• Improvements in prediction for compounds with many ionization centers.
• Greatly improved database handling and training.

Higher accuracy of apparent EXACT pK_a calculation for complex compounds with several ionization centers:

Recent changes in the pK_a calculation methodology resulted in improved accuracy of pK_a calculations for molecules with 3 and more ionization centers. The improvement is especially notable for compounds with ambiguous dissociation chains and close values of pK_a microconstants for separate ionization centers.

The new unique methodology complements the existing algorithm that is known to deliver reliable results for compounds whose pK_a microconstants for the same step of dissociation are fairly distinct (i.e. protonation can usually be determined unambiguously), and addresses the challenges of the compounds with close pK_a microconstants, as well as compounds containing a large number of ionization centers where the determination of a clear dissociated sequence might be difficult.

The new algorithm developed for ACD/pK_a DB version 9.0 results in the improved calculation of the apparent EXACT pK_a value that most closely models the measured pK_a for the compound.

This new approach also has a prominent effect on the calculation of the distribution of ionic forms and pH profiles such as solubility and logD pH curves.

New database interface and structure:

To eliminate the need for users to switch between different databases of physicochemical properties, a single unified database is introduced in version 9.0. The new pooled database enhances software training capabilities and simplifies data handling in multi-product installations such as ACD/LogD Suite.

Other improvements to the database mode include:

• New training management options improve convenience and workflow of choosing experimental data for algorithm training.
• Further integration between different ACD/PhysChem desktop programs.
• Augmented template reporting.
• Visualization of data with graphs and plots.

Added ability to calculate user-defined properties helps customers add results from their computational tools to ACD/Labs PhysChem software.

Formulas and scripts can now be added to the User Data fields to simplify data analysis.

Examples of use include:

• Marking and retrieving records where a particular value is outside the specified range.
• Retrieving records that lack information in a specific field.
• Calculating a rule by comparing different values in a number of fields to a set of criteria, and then marking the records accordingly.

Record searches can include both regular and calculated data fields.

Version 7.0 to 8.0

Extensive changes to ACD/pK_a DB, version 8.0, include significantly increased speed of calculation, added Assisted Multi-Generation (AMG) capability, a new expert tool for algorithm customization, and numerous enhancements to the interface, searching, and reporting options.

Version to version, we are continuously improving the accuracy of prediction for various chemical classes. Our recent study of prediction accuracy, conducted using a test set of >12,400 compounds, indicates that version 8.0 of ACD/pK_a DB shows continuous improvement when compared to earlier versions. Currently, 84% of calculations have a "predicted-experimental" error of less than 1 pH unit, as compared to 79% for version 6.0. View graph.

Considerably improved speed of calculation:

The calculations are now seven to ten times faster!

New AMG capability: multiple properties, multiple compounds, on the desktop:

• For small libraries and other data sets with fewer than 999 structures, you can now import and process SDfiles using the standalone desktop ACD/pK_a DB package. Calculated properties can be exported back to the SDfile, with all of the original SDfile data intact, or reported in accordance with the customizable ACD/ChemSketch templates. You can also review all of the calculation protocols one-by-one through the usual desktop ACD/Labs interface. Essential for library design.
• For customers who have several ACD/PhysChem modules, it is now possible to calculate all properties for that compound at once, view them on the screen (History Window), and export them to an SDfile.

Unsurpassed system training abilities added through NEW ACD/pK_a Accuracy Extender:

• Expert chemists can now add new Hammett equations for their novel chemical classes. These equations are automatically incorporated into the algorithms of every ACD/Labs product that references the ACD/pK_a predictions. The result is enhanced prediction for all analogs from the new class.

Augmented database search options:

• View and save all search operations and list manipulations performed on the currently open database. Store these search history macros as buttons on the Top toolbar for quick repetitive execution.
• Perform case-sensitive text searches through User Notes and User Fields.
• Search by Screen Form by either creating a search query on the screen, or using the current record as your search query.

Additional programming options for your database:

• With the newly implemented Add-on Organizer, use customized Add-ons to resolve specific challenges in your work, for example, remove the counter-ions from salts drawn with a covalent bond or in ionic form to unify input for the predictions.

Customized reports can either be generated from your single result (Results window), or use the database custom-designed Screen Form as a template (History window). You can also report all calculated properties at once from the new History Window. Print or save as ACD/ChemSketch or Adobe PDF files.

Version 6.0 to 7.0

Consistent with our commitment to continually update and improve our software products according to our user's needs and requirements, we introduced a number of new features to version 7.0 of ACD/pK_a DB.

pK_a prediction algorithm was improved for compounds with a thioxophosphine (P=S) fragment and N-containing aromatic rings resulting in the higher accuracy of calculations.

Furthermore, this new version allows you to:

• Check and choose Tautomeric forms. Prior to prediction, the user can check the existence of tautomeric forms, and choose which structure to use for prediction. The option can be switched OFF and ON.
• Report your calculations:
  • Send History file or a User Database record by e-mail from the toolbar menu.
  • Make reports at a click of a button. Select the data to send to the report and use the ACD/ChemSketch reporting tools to create a finished document.

Numerous helpful features are added to help create and manage User Databases:

• Avoid mistakes when entering new data into a User Database. You may choose to display a warning if the difference between the entered value and the calculated value is greater than a pre-determined variation set by the user.
• Compare databases, merge them, or find common intersecting records in the User Databases. Track the success of the merging operation with the Merge Information log.
• Search by multiple User Data fields simultaneously. Boolean logic operators such as 'AND' and 'OR', or their combination, allow users to perform advanced searches.
• Highlight the searched substructure. Visually identify the fragments of the molecule that were requested in the original query. Choose the color of display and switch the option on and off.
• Search the database for ALL tautomeric forms. Software checks if the structure is available in different tautomeric forms and suggests the search options:
• Enhanced SDfile import/export. Allows user to preview the field names in an SDfile during import to allow mapping to existing fields in the database and Notes subwindow.

Version 5.0 to 6.0

ACD/pK_a DB, v. 6.0, features improved prediction and training capabilities and offers a number of new exciting options:

• Similarity Search: in addition to the existing structure and sub-structure searches, you can now search the database to discover compounds with similar chemical structures;
• Expanded internal database almost DOUBLED in size due to the addition of data provided by BioByte Corp. Now the database contains 15,932 compounds, as compared to 8,882 compounds in version 5.0.
• Enhanced accuracy of prediction as a result of the pK_a calculation algorithm improvement for several classes of compounds. pK_a training dataset was substantially increased due to the addition of BioByte Masterfile data;
• Enhanced software training with user's data: pK_a prediction engine can now be simultaneously trained with multiple databases to improve the accuracy of calculations for user's datasets and to reduce training time;
• ACD/pK_a DB database can now be viewed as a table.
• SDfile Import/Export options can be customized.