4th Annual Cell Based Assays 
Main Conference Day Two
Wednesday, October 8, 2008
| 8:30 | Morning Coffee |
| 9:00 | Chairpersons’ Recap of Day One - Shalini Gupta, PhD, Director, Medical Sciences, AMGEN |
| 9:15 | Statistical Process Control for Improved Assay Development Production and Transfer The most important aspect of how we develop, validate, maintain, and possibly transfer assays is rooted in the data. It is the data that inform us as to how we should act, where we should go, and if our processes are capable of performing at some required level. Data analysis is the key to this and statistical techniques are tools that we use. Discusses types of charts that are used, how these charts can tell us if our processes are in control, and where we can use these charts in the development, production, and transfer processes. - Martin Kane, MS CRE, Senior Manager, Process Statistics, HUMAN GENOME SCIENCES |
| 10:00 | Optimization and Analysis of Sources of Variation in theWestern Blot Neuro-2a cell based Assay Via Application of the Weighted Four-Parameter Logistic Model Purpose. The Neuro-2a cell based assay is a potency assay that uses 96-well plates and differentiated Neuro-2a cells in a Western Blot read-out. It has been designed to measure all three steps in BoNT/A activity: receptor binding, internalization and translocation, and catalytic activity. Mathematical modeling of the Neuro-2a dose-response curve via the four-parameter logistic model [4PL] has been used not only (1) to improve this assay but also (2) to characterize it after finalization. Methods: (1) The 4PL was applied to actual data and then used to simulate the likely results for three different dose regimens of interest. The 4PL was then applied to the simulated data and the resulting model parameters and convergence were monitored. The optimal dose regimen was defined as the one yielding estimated parameters as close to nominal as possible while also causing the fewest convergence problems. (2) After optimization, routine testing was initiated, yielding a historical database spanning nine months. Weighted four-parameter logistic models were separately fitted to the individual response curve data in this historical database. The fitted asymptotes, EC50, HillSlope, and weighting parameters values were then regressed on variables measuring the relevant extrinsic variability sources (such as operator id and cell passage numbers). Results: (1) The nature of the collaborative relationship, and its benefits, between the Depts. of Biostatistics and Biological Sciences will be discussed. Appropriate dose optimization was a direct result of this collaborative effort. (2) Statistically significant sources of extrinsic assay variability will be noted and quantified. The results will then be used to indicate how a unified mixed-effects weighted four-parameter logistic model might be built. The benefits of unified mixedeffects model-building will be noted. - David Paul, PhD, Manager, Department of Biostatistics, ALLERGAN |
| 10:45 | Morning Networking Break |
| 11:15 | Statistical Analysis of Cell Based Assay Data
- David Paul, Manager Biostatistician, Biostatistics, ALLERGAN |
| 12:00 | Networking Luncheon |
| 1:15 | Topic TBA Pankaj Oberoi, Meso Scale Discovery |
| 2:00 | New Technologies Applied to Cell Based Assay Development
- Pankaj Oberoi, Mesoscale Discovery - Stephen Ullrich, Senior Scientist, HUMAN GENOME SCIENCES |
| 3:00 | Conference Concludes |
Bottom menu
Copyright 2003-2006 IIR Holdings, Ltd. All rights Reserved