1. You should specify what you add to your in vitro tests. You can add one regulatory protein, or none, or both; you should specify which ones are added.

2. It's not enough to say "you can isolate a super-repressor" or whatever; there has to be a specific and plausible molecular mechanism for doing so.

3. When you describe a particular experiment or set of experiments, say how the models predict different outcomes.

4. For the biochemistry part, it's often easier to compare the models by inspection and see how they differ, as opposed to making exhaustive lists of their predictions. I don't want these lists anyway, but you may be able to avoid the effort of making them by comparing the models.

5. Please make clear how you are distinguishing the models--it is not enough to give lists of predictions!

Ways people went wrong on the biochem part:

1. It is important to give the prediction first, not the experiment. State the prediction, then say how you will test it.

2. It is crucial to compare the predictions for the same experiment. As an absurd example, an answer of the following form is not correct:
"Model B predicts that the target gene will not be expressed in the absence of the inducer, while model H predicts that the target gene will be expressed in the presence of the inducer." Clearly, while both statements are true, they are also true of the opposite model. A more realistic version would be "Model B predicts that Regulator I binds to the regulator II gene in the absence of the inducer, while model H predicts that regulator I is a protease". Both are true, but they should be expressed in two separate predictions, as in "1. B predicts xxx, H does not. 2. H predicts yyy, B does not."

3. You need to be clear about how to decide which one is Regulator II. In the sample answer it is defined as the protein that binds the target gene. You could also define it with an epistasis test (for these three models anyway). If in model H you call the protease regulator II, then compare it to Reg II from B or D you're bound to get messed up.

4. In the ground rules for this exercise, you have two purified regulatory proteins; regulator II is defined as above. Importantly, you don't have four proteins, two for each model! You have two proteins, and the biochemistry experiments are designed to analyze their properties.

5. In the in vitro transcription experiment, you need to have a control without the DNA-binding regulatory protein. Otherwise you can't tell how this protein is affecting the level of transcription. Even in the absence of an activator or the presence of a repressor, you'll get some transcription (e.g. see my LexA paper); you need to compare this with the signal you get in the presence of activator or absence or repressor. See here for more details.

6. You can't get translation of proteins in an in vitro transcription system. Recall that this contains RNA polymerase, DNA, buffers, and perhaps your regulatory protein. Think about what the cell needs to carry out protein sythesis.

http://www.biochem.arizona.edu/classes/bioc568/bioc568.htm
Last modified October 29, 2007
All contents copyright © 2007 All rights reserved.