# Talk:Stochastic models of ion channel gating

Thanks for the helpful review. Our comments are in bold. -- TK & JW

it would good to give an idea of typical densities of channels such as the potassium channels in the Morris-Lecar model - one could dig up the original values from the experimental work. This would give a better idea of whether 1000 channels per unit area is still small or close to the real thing. Of course this answer will differ across channel type and systems, but it would be useful to get an idea in a specific case.

We have made this calculation, and included the results in the section Is Channel Noise Important? For the barnacle giant muscle fiber, we estimate over 100,000 potassium channels, and point out that channel noise is not likely to be important. We also discuss a case in which we estimate the number of channels to be 1000-4000, for which we have demonstrated directly the importance of channel noise.

I also think that the equations for the modified ML (with noise on the recovery variable) should be put in the text. Maybe it will be sufficient to put it in the matlab code link, where the equations are spelled out, but I can not find anything there now.

We have added the equations and fixed the link to the MATLAB source code.

In Approximate Methods, you mention that the rate of escape is progressively slower. The forward and backward rates are slower when you go to the left of the plot. So you should clarify: escape from what?

We have clarified this point.

Also, the connection of the Markov scheme to the history-dependent scheme is a bit abrupt. Gillespie can simulate Markov as well as history-dependent schemes (for example it can include non-exponential waiting times). Here you write that the history-dependent scheme can be approximated by a Gillespie algorithm applied to a Markov chain. So the physical approximation is really the Markov chain, and Gillespie is the method used in the cited reference to simulate that chain. So a few more words would be useful to clarify the link.

We have clarified this point as suggested.

Finally, the last paragraph mentions flicker, which is not defined, although pretty obvious. Better to define it as stochastic open-closed transitions.

We have made the requested change.