| Abstract: |
Analytical ultracentrifugation is a first principle method that gives the sedimentation and diffusion coefficient of molecules of various sizes and shapes in the liquid phase. LAMM equation which is a partial differential equation and is solved numerically using the finite element methods explains the spatial and temporal changes of molecules under centrifugal forces. Optimization methods like the least-squares approach are used to develop a mathematical model that best fits the experimental data using the LAMM equation. This is a theoretically challenging subject called inverse problem and in most cases, it needs high-performance computing to give a result. The problem gets even more complicated when we see that there are lots of noises from different sources in each measurement when we work on the real samples. So the goal of data modeling is to develop the best model and find the noises from the experimental data and minimize them in the final model. There are three different types of time-invariant noise, radial-invariant, and stochastic noise, which only the last one cannot be identified in the data modeling procedure. Fortunately, there is a robust approach to minimizing the time and radially-invariant noises in the sedimentation velocity experiments. But what about the equilibrium experiments? The radially invariant noise can be neglected in these experiments because the last scan in which the system has already reached the equilibrium point is taken for further analysis. But, time-invariant noise still remains, and current approaches cannot deal with this type of noise. In this study, we present a way to capture the fingerprint of Optima's optical system at different wavelengths, which are believed to be the main source of time-invariant noise. We have tested this idea by calculating the pseudo-absorbance profile of water molecules at different wavelengths and time intervals. |
|---|
