Comparing Simulation with Experimental
Tweaking simulation was not bringing it any closer to the experimental shown in the last post. This led me to try the experimental once more. I set up the same system and used 1:2 dilution of the standard protein mixture. I then homogenized the retentate and filtrate and ran 10 ul of each on a gel:
One concern I had with this is that the resulting coomassie stained gel seemed to have some saturation in the retentate bands. A saturated retentate band would lead to a lower estimate of the concentration and thus a higher ratio. I reran this gel with only 5ul of each sample:
This gel had much less saturation, but there may still be some. However I may not be able to drop the size of the starting sample much more without losing the signal from the filtrate.
Anyway, densitometry of the new gel leads to this figure:
It’s clear that the simulation matches much better with the experimental now. Some possible reasons for the differences:
-still unclear if there may be saturation/error in the densitometry
-all size measurements of molecules roughly taken from xtal structure
-diffusion coefficients calculated from those sizes
-effects of protein shape/stickiness/adsorption