 |
| Dr. Antony J. Williams |
Periodic Features < ISSUE #2
October 1999 |
ACD Toolset and the Interactive Laboratory
Advanced Chemistry Development specializes in PC and Web-based software for structure drawing, 1D and 2D NMR processing and prediction, databases of H-1, C-13, F-19 and P-31 chemical shifts, processing for UV-Vis, IR, MS and LC-MS data, chromatography, spectroscopic and chromatography database management for NMR, IR, MS and UV-Vis; chemical property prediction including Boiling Point, pKa, logP, logD, and solubility; and IUPAC and CAS Index systematic chemical naming, web-based LIMS, and integration with MDL/ISIS and Cambridgesoft software. The intention of this short article is to present an overview of some of these tools and an examination of their unique capabilities. Also, the integration of these prediction tools with an online service delivers the ACD/I-Lab, the Interactive Laboratory.
 The original intention behind ACD/I-Lab was to provide global access to the industry standard ACD prediction algorithms using client-server technology and to develop cross-platform, Web-based client software. In the process, we demonstrated the utility of Java in the world of Chemical Information by using Java applets for structure drawing and results display. We also offered a route for third parties to reuse the existing Web interface and client databases for their own prediction engines or database searches.
The resulting application provided a universal Web-based gateway to various chemical information resources, property prediction programs and chemical databases utilizing Java-based structure drawing and spectral display applets to provide structure submissions for prediction purposes. ACD/I-Lab was unveiled to the public in September 1996 with an initial offering of CNMR and HNMR prediction only. By the summer of 1997, we averaged over 400 predictions per day and we expanded capabilities to include the following database searches and property predictions: IUPAC name generation, CAS name generation, HNMR spectrum prediction, CNMR spectrum prediction, pKa database search (over 9000 structures), pKa prediction, logP prediction, boiling point prediction, vapor pressure prediction. As of March 1998, we had over 3500 registered users and had performed over 90,000 predictions and database lookups.
|
| The integrated Chemsketch 3D viewer
|
We are now proud to announce the arrival of ACD/ILab-II. In this case, the prediction algorithms and database lookups have been directly integrated into ACD/Chemsketch, the FREE full-featured structure drawing package for Windows which has generated over 22,000 downloads since June 1998 (http://www.acdlabs.com/download/). An online module allows the ChemSketch user to interact with the ILab server and appropriate helper applications are automatically started to view results such as predicted spectra. ChemSketch is also integrated with ACD/ChemBasic, the ACD Chemical 'Beginners All-purpose Symbolic Instruction Code'. It allows the customization of ACD software, or connection of your own programs to the ACD interface. It is a true macroprogramming language, with common statements and functions for dealing with numbers or texts; also, built-in molecular objects and functions to manipulate them. A ChemBasic macro can, for example, get a molecule from the ACD/ChemSketch window, calculate its logP, replace one fragment of it with another and then recalculate logP. Or, a macro can get a shorthand formula, e.g. of an oligosaccharide like GalNAc(alpha1-3)Gal(beta1-4)[Fuc(alpha1-3)]Glc(betaOMe) convert it into a ChemSketch diagram, perform conformational analysis and depict the most probable 3D-structures. ChemBasic, and many sample macros, can be downloaded as freeware from our website.
There have been previous overviews of the original ACD/ILab so here I will review solely the new aspects of ACD/ILab-II and the prediction capabilities of our standalone Windows-based products. The PC-based prediction modules have been made available through other popular structure drawing modules including MDL/ISIS and Cambridgesoft's ChemDraw using add-in extension capabilities. The ACD/Labs Extension for ChemDraw is available at no charge from our FREE STUFF link at the website and include a number of free prediction modules. When other ACD modules are installed ChemDraw simply becomes the drawing interface of choice while the ACD display interface can be used for displaying results.
Naming Modules
|
| What's in a name?
|
IUPAC Name Pro generates the systematic chemical name according to IUPAC rules for a drawn chemical structure. IUPAC Name Pro shows how the parts of the name relate to each structure fragment, and displays a hyper-linked text of the IUPAC rules, to explain nomenclature conventions. A wide variety of name preference settings for preferred stems, retained names and so on are available. More than 40,000 functional groups and all non-noble-gas elements are supported. Organometallic compounds with sigma carbon-metal bonds; substituted amino acids; retained names for amino acids and monosaccharides; fused systems; biphenyls; heterochains and heterocycles plus many, many more possibilities are available. Stereochemical bond designation allows you to name Z/E and R/S isomers in a more complete fashion than any other commercially available naming software available today. We are official members of the IUPAC committee for naming as a result of our expertise in this area.
Index Name Pro generates systematic chemical names according to Chemical Abstract Services (CAS) Index rules using all of the interface strength of IUPAC Name Pro. For a competitive analysis of our software with other commercial packages visit our website.
NMR Prediction
H-1, C-13, F-19, P-31 NMR Predictors
|
| 13C predicted spectrum
|
Each of these predictors is a separate, yet integrated module, which can be used for the prediction of NMR spectra for input organic structures. NMR spectra can be calculated for almost any drawn organic structure, or a mixture of compounds, to an accuracy of ± 3ppm or better for C-13 or ± 0.3 ppm for H-1 in most cases. The calculated chemical shifts are provided with the confidence limits. The predictors are based on large internal databases with a great number of experimental chemical shifts for molecules, and a unique fragment-based algorithm to quantitate precisely the intramolecular interactions in new organic structures. The predictors calculate the spin-spin interactions of the nuclei with the magnetic nuclei of other elements, proceeding from the natural ratio of magnetic isotopes. They recognize spectral differences between the following types of isomeric structures: cis-trans isomers of alkenes; cis-trans and endo-exo isomers of cyclic structures; and syn-anti isomers of amides, oximes, hydrazones, and nitrosoamines. For H-1 and C-13, the calculated spectra can be expanded in any region, and easily attribute spectrum peaks and chemical shifts to the nuclei and vice versa.
|
| Substructure search of the C NMR database
|
The Calculation Protocol window allows direct examination, as a histogram plot display, on a nucleus by nucleus basis, of which structures within the database were used for the predictions. In this way all database structures utilized for the prediction of a single molecule can be screened one nuclear centre at a time. H-1 and C-13 ACD/NMR Predictors simulate operation of an actual spectrometer, allowing you to choose among different modes of obtaining spectra (off-resonance, DEPT, J-modulation), operating frequency, solvent, and concentration of solute(s).
|
| Expansion of proton spectrum including shifts and coupling constants
|
Both the H-1 and C-13 predictors contain a 'self-training' system which increases the accuracy of the prediction. You can create and/or update your own database with new experimental or literature chemical shifts. The next time you open your database and make calculations for any new compound, the program automatically uses the information from both your opened database as well as its own internal databases, so that the accuracy of calculations is increased. Add-on databases are available with the following numbers of structures: H1>82,000, C13>67,000, Natural Products C13>6,000 structures, F19>11,500, P31>18,500. These databases include with original references and can be searched directly by structure, substructure, molecular formula, molecular weight, IUPAC name, chemical shift value or values and coupling constant(s).
Phys-Chem Properties
ACD/Boiling Point & Vapor Pressure is a program to calculate the boiling point for a compound, from 0.001 torr to 10 atm. Our predictions are normally accurate to within ± 5ºC for structures with fewer than two polar groups, and to within ± 10º for structures with more polar groups. It will also calculate the flash point, and vapor pressure as a function of temperature, and will calculate the enthalpy of vaporization at the boiling point. Select the graph and table output you desire: boiling point vs. pressure or vapor pressure vs. temperature. Select the units for these, and also for the enthalpy of vaporization. You can choose to print out any or all of this information. The flash point Protocol window shows you how the value was estimated. To determine constants and relationships, our algorithm uses an internal database which contains boiling points for 10,000 compounds — 6000 of which have a boiling point at 760 mmHg. Additive increments of atoms and groups for our function K and force increments of interactions between groups and atoms were calculated from this database. Your compound does not have to be in the database in order to have its boiling or flash point calculated.
ACD/LogP DB gives a quantitative structure-based estimation of the logP partition coefficient and its 95% confidence interval at the pH for which the neutral form of the molecule exists. It also gives the aqueous solubility (S), the bioconcentration factor (BCF); and the adsorption coefficient (Koc) for molecules containing H, B, C, N, O, F, Cl, Br, I, Si, P, S, Ge, As, Se, Sn, or Pb. It contains a searchable database with thousands of chemical structures with original references. In most cases, ACD/LogP calculates logP values with an accuracy ± 0.3 or better. It automatically checks the alternate ionic and/or tautomeric forms of the drawn structure and warns you that you may encounter difficulties in measuring the accurate logP value experimentally. It also provides a detailed explanation window which describes how the logP value was calculated. It exports and imports SDF files, and, if an SDF file contains chemical structures without logP values, the program can calculate them automatically during the process of importing structures. The unique ability exists to train a user database with either experimental logP or logD values and our unique fragmentation algorithm will generate user fragments and utilize them in later predictions for related structures.
|
| One of the multiple pulldown menus of property predictions
|
ACD/pKa DB calculates the acid-base ionization constant, pKa. Each calculation is provided with its ± 95% confidence limits and literature references, if available. The accuracy of calculations for simple structures is usually better than ± 0.2 pKa units. The predicted pKa is based on the choice of the dominant ionic form in the system at the equilibrium state under standard conditions (25º C and zero ionic strength). ACD/pKa shows the ionization center for each pKa value, and can also calculate the microscopic constants of dissociation of specific ionic forms which could be present in an equilibrium system. ACD/pKa DB includes a fully searchable database containing more than 8,800 chemical structures with references to more than 23,000 experimental pKa values.
|
| Windows-based logD prediction using integrated logP and pKa prediction
|
The ACD/LogD Suite calculates the octanol-water partition coefficient for almost any drawn organic structure at any pH. Besides the convenient tools for the quick estimation of logD at any pH, it calculates logD with or without ion-pair partitioning; the percentage of the dominant form in aqueous and organic phases; the aqueous solubility; the bioconcentration factor (BCF); and the adsorption coefficient (KOC) at any pH (0-14). It will show the dominant ionic form of the chemical structure in the water-octanol mixture, and can present results both in table and graph form.
Additionally, ACD/LogD Suite includes access to the ACD original pKa and logP databases and prediction of these properties. Batch versions to process 1000s of structures at one time are available for ACD/LogP, pKa and LogD software.
Now ACD/I-Lab-II allows access to the strength of each of these prediction modules online using ChemSketch online. When the user has logged in a number of menus become available for systematic naming, physicochemical property prediction, NMR prediction and database searching by structure and substructure. Some example screen shots are shown below.
Download ChemSketch online today from http://www.acdlabs.com/ and try out the power of ACD/I-Lab II.
Go back to the top
|
|
|
|
