|
|
PITTCON® 2005 Poster Schedule
Orlando, FL, USA
Feb. 27 - Mar. 4, 2005
Orange County Convention Center
North/South Complex
Orlando, FL, USA
|
|
|
|
View Talk Schedule
Poster Schedule
| |
| Title: | Mutual Automated Peak Matching in Chromatographic Method Development |
| Authors: | Mike McBrien, Ed Kolovanov, Andrey Bogomolov, and Vitaly Lashin (ACD/Labs) |
| Date: | Tuesday, March 1, 2005, from 9:30 AM - 12:30 PM |
| Abstract #: | 810 - 25P |
| Abstract: | Computer-based optimization tools streamline the process of chromatographic method development considerably. Peak movement is modeled as a function of one or more variables, allowing for effective visualization and optimization of the experiment across all potential parameter values. The speed and ease of application of tools are limited, however, by the necessity of tracking the location of each analyte across each of the chromatographic runs.
This paper will describe new algorithms designed to take advantage of hyphenated techniques such as LC/UV-Vis and LC/MS in order to automate the chromatographic peak matching process. The principles of Mutual Automated Peak Matching (MAP) will be discussed as well as performance of the algorithm on real samples under conditions of varying signal/noise ratios and solvent conditions.
|
|
| Title: | Starting Points in Chromatographic Method Development - a Structure-based Approach |
| Authors: | Dr. David S. Bell, Mr. John Martin, and Mr. Hugh M. Cramer (Supelco Division of Sigma Aldrich) and Mr. Eduard Kolovanov, Mr. Alexey Danilov, and Mr. Mike McBrien (ACD/Labs) |
| Date: | Tuesday, March 01, 2005, from 9:30 AM - 12:30 PM |
| Abstract #: | 810 - 3P |
| Abstract: | The initial parameters chosen for a given method development task often determine the extent of time required for the study as well as the quality of the final conditions attained. Frequently the method developer unknowingly travels down dead-end paths. With continually shrinking timelines allowed for method development studies, poorly chosen starting points can lead to missed deadlines or hastened completion of the task. The latter of which often generates adverse issues later in the life cycle of the method.
One of the largest challenges in chromatographic method development lies in determining where to start experimentation. A number of concepts have been applied to this problem, including generic starting points, structure-based searches of chromatographic application databases, and physicochemical prediction. Utilizing new software technology, it is now possible to use a combination of all three techniques to effectively select between method development starting points.
Software tools have recently been designed to accurately predict retention times for compounds in a sample based on a previously-collected knowledge base of experimental retention times for compounds in a set of generic "starting point" experiments. Multiple complementary methods have been designed that target very different selectivities such that there is a maximal chance that any given set of compounds will show some resolution. Thousands of chromatographic retention times form the knowledge base for prediction of retention times of new compounds with very high accuracy. The experimental design and application of the experiments will be discussed in this paper.
|
|
| Title: | Spectrum: Structure Correlations in Infrared and Raman Spectroscopy, A Corporate Resource |
| Authors: | Michael Boruta, Michel Hachey, and Antony Williams (ACD/Labs) |
| Date: | Wednesday, March 2, 2005 from 1:30 - 4:30 PM |
| Abstract #: | 1620 - 16P |
| Abstract: | Many industrial and research laboratories spend a significant amount of time, effort, and expertise to understand specific classes of compounds. This knowledge can be of critical value to a corporation. Although the value of this knowledge is recognized, its ability to be stored, transferred to new researchers, and shared with other potential users has been limited, hence limiting its utility as a corporate resource.
This paper will examine some of the current practices in use and their limitations. We will also discuss some new methods that can be used to obtain, store, and retrieve this knowledge so that, once derived, it will remain a resource within the corporate laboratory beyond the confines of individual analyst expertise. The goal is to not only capture your own knowledge but also to combine it with that of your colleagues. This will allow you to make better decisions and give you the ability to share this knowledge with current and future colleagues.
|
|
| Title: | Software-Assisted Selection of Generic Separation Methods in the High-Throughput Laboratory |
| Authors: | Mark Woodruff and Charlotte Blythe (Thermo Electron Corporation, Runcorn UK), Margaret Antler, Alexey Danilov, Rhiannon Jones, and Mike McBrien (ACD/Labs) |
| Date: | Thursday, March 3, 2005 from 9:30 AM - 12:30 PM |
| Abstract #: | 1870 - 20P |
| Abstract: | Generic chromatographic methods are generally a small set of separation methods that are designed to produce sufficient resolution for the majority of samples in a situation where it is not practical to spend time developing high quality methods for specific samples. High-throughput and walk-up laboratories thus rely on generic or standard separation methods for structure verification and purity estimation. Software tools can further increase sample throughput by evaluating which method in the set of generic methods will be most appropriate for a particular group of compounds. In addition, data quality can be increased by ensuring that compounds are retained sufficiently on the column and/or can be expected to show resolution from expected/unexpected impurities. The software works in the following manner. For a particular set of methods, the software is first trained. A number of representative samples are analyzed using the set of generic methods, and the results are entered along with their chemical structures into the software. Once the initial training is complete, the chemical structure(s) of the novel compound(s) are entered into the database. A structure-based retention model is developed for each generic method using the most similar compounds in the database. Selection between each of the candidate methods is done based on the predicted results.
This paper will describe the design of typical generic methods, the column selectivity required, and the selection of compounds for the training set. Results will then be shown for some "unknowns", highlighting how the correct choice of above criteria for the training set leads to excellent prediction capabilities from the software.
|
|
| TOP |
This page was last updated
12 September 2007
|
