|
|
PITTCON® 2006
Orlando, FL, USA
March 12 - 17, 2006
Booth #: 2340 - See map for location of booth
|
|
|
|
ACD/Labs' Talk Schedule
| |
| Title: | Advances in the Extraction of Potential Metabolites Using a Self-Optimizing Componentization Algorithm for Peak Extraction and Identification of MS1 Datasets |
| Authors: | Robyn Rourick, Mark Bayliss, Vitaly Lashin |
| Date: | Tuesday, March 14, 2006 @ 10:05 AM |
| Session Title: | Advances in LC-MS Strategies for the Identification of Impurities, Degradants, and Metabolites |
| Location: | Room 208A |
| Abstract #: | 830-5 |
| Abstract: |
Highly automated methods featuring fast chromatographic separations along with mass spectrometry analysis have become the benchmark analysis strategy for metabolite identification. However, the most significant bottleneck in support of metabolism studies resides in the data interpretation for determining metabolic liabilities. Various chemometric approaches for automated peak extraction have been proposed over the years, included are algorithms such as MEND, CODA, and COMPARELCMS. Each has advantages and disadvantages in the fields of noise reduction, peak extraction, and dataset comparisons. These approaches have proven valuable for metabolic profiling but have required extensive understanding of the operational parameters to produce satisfactory levels of output fit for a routine environment.
Chemometrics in metabolism studies has always been challenging due to the complex matrices that are common in these samples. Significant baseline disturbances, due to solvents and the sample matrix, challenge these algorithms to differentiate chromatographic peaks from noise. Whilst the application of MS/MS makes a significant reduction in the presence of matrix, a number of scanning techniques and instrument types are often required to gain access to the broad array of information that must be used to determine a structure.
This presentation will focus on a combination software approach involving chemometrics and mass spectral knowledge for the extraction of potential metabolites and auto-identification of molecular ions for chromatographic components. These algorithms are able to extract metabolites by searching for semi-quantitative differences within a series of compared datasets through self optimizing data extraction parameters. Early applications suggest potential as a useful tool for screening metabolite samples in a relatively facile manner, with less dependency on acquisition methods and instrument platforms.
|
|
| Title: | Chromatographic Peak Tracking Through Chemometric Analysis of Hyphenated Data |
| Authors: | Michael McBrien, Alexey Galin, Vitaly Lashin, Eduard Kolovanov, Mark Bayliss |
| Date: | Tuesday, March 14, 2006 @ 10:25 AM |
| Session Title: | Chemometric Applications |
| Location: | Room 202C |
| Abstract #: | 860-6 |
| Abstract: |
Computer-based optimization tools have streamlined the process of chromatographic method development a great deal by eliminating trial and error, and instead enabling systematic investigation of chromatographic systems. These tools reduce the number of injections required to optimize a given chromatographic system for a given sample, and increase the capacity for design of methods that meet simultaneous scientific criteria of resolution, run time, and robustness. One of the only drawbacks of computer-assisted method development is that chromatographic peaks must be tracked accurately from one experiment to the next, often requiring a great deal of time spent on manual interpretation of the chromatographic data. The advent of hyphenated chromatographic detection in non-routine analytical laboratories has created an opportunity for the application of new automated techniques for interpretation of the data associated with method development.
This talk will discuss the development of algorithms to take advantage of diode array detection and mass spectrometry in order to automate the chromatographic peak matching process. Two chemometric algorithms will be described which exploit the benefits of their respective detection systems, along with algorithms to reconcile the results of each. Examples of applications of this tool, dubbed Mutual Automated Peak matching, or MAP, will be discussed. The role of the specific detectors will also be discussed, as well as limits of detection and isomer differentiation, and sensitivity to changes in experimental conditions.
|
|
| Title: | Addressing Data Management Needs in a Preformulation Environment |
| Authors: | Michael Boruta, Michel Hachey |
| Date: | Tuesday, March 14, 2006 @ 2:30 PM |
| Session Title: | Laboratory Informatics II |
| Location: | Room 204A |
| Abstract #: | 1190-4 |
| Abstract: |
Salt and polymorph screening are performed at an early stage of the preformulation process, between the late drug discovery and early development stages. The early salt form choices have far-reaching effects on the ultimate efficacy, ease of manufacturing, and profitability of the final active pharmaceutical ingredient (API) product formulation.
The decision on which form to pursue is based on numerous experiments designed to explore the chemical and physiochemical properties of the different forms and methods to produce those forms. High throughput (HT) technology is being used to increase the chances of finding suitable forms for development, creating even greater amounts of experimental data to analyze.
In this work, we discuss the acceleration of the API salt form selection through state-of-the-art analytical data management systems. With such systems, real-time analytical data obtained from disparate instrument types can be integrated into a single user-friendly interface. An Analytical Data Management System (ADMS) solution can provide end users with significant capability for visualizing, processing, and comparing data across large groups of experiments, ultimately facilitating decision making.
|
|
| Title: | A Fully Automated System for Chromatographic Method Development Utilizing LC/MS/DAD Detection |
| Authors: | Michael McBrien, Andrey Vazhentsev, Vitaly Lashin, Alexey Galin, Eduard Kolovanov, Mark Bayliss |
| Date: | Wednesday, March 15, 2006 @ 10:25 AM |
| Session Title: | Pharmaceutical Analysis: Automation of the LC Method Development Process |
| Location: | Room 203A |
| Abstract #: | 1400-6 |
| Abstract: |
The development of chromatographic methods can be a very time-consuming undertaking, particularly for complex samples. Computer-assisted method development has helped to lessen this problem, enabling chromatographers to typically create more effective methods in less time. However, there is still an onus on the researcher to interpret the output of the optimizations, and to input subsequent experimental parameters to the instrument. The next logical step in chromatographic method development is to connect method optimization software directly to instrument software, enabling the system to investigate the experimental "surface" without the necessity for manual intervention at regular intervals.
This paper will describe the application of hyphenated detection (mass spectrometry and diode array detection) combined with control of chromatographic data systems to form a complete automated method development system. The automation of data interpretation and subsequent instrument control creates the opportunity for extremely rigorous method development, including possibilities for new approach and evaluation strategies as well as visualization and data organization tools. The underlying techniques and philosophies of chromatographic method development will be discussed in the context of the application of the technology to specific problems in method development.
|
|
| Title: | Is it Possible for Software to Automatically Determine the Molecular Ion for Each Chromatographic Peak in an LC/MS Dataset and Automatically Calculate Appropriate Extraction Parameters? |
| Authors: | Mark Bayliss, Vitaly Lashin |
| Date: | Thursday, March 16, 2006 @ 1:30 PM |
| Session Title: | HPLC: Method Development |
| Location: | Room 205B |
| Abstract #: | 2320-1 |
| Abstract: |
Previously, chemometric algorithms have been developed for the extraction of chromatographic peaks from LC/MS datasets. It is often recognized that these software driven extraction approaches can significantly reduce the processing time for scientists. Historically it has not been possible for software to automatically determine the molecular ion for each eluting peak and be able to reliably determine optimal processing parameters on a per dataset basis, hence removing the need for often complex setup and optimization by a user.
Automated confirmation software systems have been developed by numerous instrument and software vendors using user-driven targeted extracted ion mass chromatograms. These systems assume that if a chromatographic peak is detected for a target mass value then this is consistent with the presence of a required chemical entity, as is often applied during chemical synthesis confirmation approaches. It is estimated that up to 10% of all automated ion confirmations using this approach for nominal mass determination may be incorrect being the result of isotopic contributions from other eluting chromatographic components. This presentation will detail a development project which is designed to fully componentize LC/MS datasets using a combination of chemometric algorithms and mass spectral ion knowledge. The output from this extensive analysis is a reduced dataset that details the molecular ions present in the dataset, 12C and 13C classifications, adduct ions, multimers and potential fragment ions. Additionally, we will discuss the challenges and the solutions framework that has been developed to facilitate this process.
|
ACD/Labs' Poster Schedule |
| Title: | Managing Data in LC/MS/DAD Forced Degradation Studies |
| Authors: | Michael McBrien, Andrey Vazhentsev, Eduard Kolovanov, Vadim Tashlitsky, Mark Bayliss, Willy Janssens, Achille Pluym, Rudy Sneyers |
| Date: | Monday, March 13, 2006 - Morning Session |
| Session Title: | Techniques for Pharmaceutical Analysis |
| Location: | Exhibit Hall |
| Abstract #: | 470-67P |
| Abstract: |
Today's chromatographers have an unprecedented capacity for generation and collection of data associated with unknown samples. The combination of immense hard drives, multiplexed detection, and robust instrumentation provides the analyst with a tremendous amount of data. In many cases, this has simply moved the bottleneck in analysis from data collection to data interpretation.
One of the largest challenges facing modern chromatographers is the problem of development of methods for detection and quantitation of impurities in drug samples associated with stability studies. A large number of potential impurities can be generated through various reactive conditions, including exposure to UV light, acid, base, peroxide, etc. Chromatographers will then design chromatographic conditions for quantitation of each of the impurities, injecting multiple related samples for each set of conditions individually. The detected components are used collectively in method development. This eases the burden of detection of low-level species, and of chromatographic peak tracking across sets of conditions. Thus multiple, relatively simple samples represent a single, more complex sample for purposes of generation of a method that will separate all of the components.
The result of combining this approach with a systematic strategy of investigating the chromatographic response of a complex sample can be very rigorous method development. However, the amount of data that is generated, and the time required to analyze it, can be onerous. Chromatographic peak tracking and cross-sample component resolution, combined with the necessity of transcribing peak tables into modeling software can require as much or even more time than the data collection itself. This poster will describe the application of new techniques for management of method development data, both for complex and "monotone" samples and the utilization of this data for purposes of rigorous method development.
|
|
| TOP |
This page was last updated
12 September 2007
|
