No on this graph

5 posts / 0 new
Last post
samita
samita's picture
No on this graph

Can some one tell me what the meaning of these no. written on the graph. Its a single channel open time graph.
If some one can understand it please explain me,
I shall be very thankful to you for your kind reply.
best regards

Fraser Moss
Fraser Moss's picture
The histogram is best fit

The histogram is best fit with the sum of two Gaussian fits.
There are two components which each has a mean. These means are the 3.4ms and 9.4ms numbers above the graph

The area under each component describes the percentage of all opening events described by that component.

So 43% of all openings have a mean opening time of 3.4ms and 57% have a mean open-time of 9.4ms.

I hope that has helped you.

samita
samita's picture
many thanks for your help.

many thanks for your help. Its really a wonderful explanation.
I am completely new to this sort of work and i need lot of help from this forum to make my concepts clear.
many thanks for your help.
regards
samita

samita
samita's picture
But can you explain me why

But can you explain me why its necessary to calculate the two exponentials for the open probability time of the channel.
What one can loose by canculating just one exponentail function.

Fraser Moss
Fraser Moss's picture
Firstly I should say that in

Firstly I should say that in my initial reply I said they had fit with Gaussians. It actually looks like Log-exponential fits, but the same points apply reagrding the two compoenets and the fraction that each represents.

It may not be neccessary to fit with two components and one could be sufficient.

When analyzing your data there are many factors you need to consider, which included selecting the appropriate bin size for the amount of data. There are many articles available online and in journals about how to bin your data correctly.

This can strongly influence whether something is best fit by 1, 2 or more components. When fitting a histogram the fitting procedure will calulate the Chi-Squared value for Goodness of fit. 0 being no fit at all and 1 being the best possible fit. The person analyzing the data will have to decide the criteria at which they are going to say the fit is good. Chi^2 = 0.90 for example or maybe 0.99. But then for all other equivalent comparisons, this value is the cut off.

It is always possible to fit biological data to more and more curves to improve the goodness of fit, but there has to be a point at which you ecide if you are introducing experimental/analytical artifact or if it is a true biological phenomena you are observing.

In the example you provide, you will have to look at the criteria they used to decide the best fit. The histogram does look like it could fit well to one component. You would have to look and see if the statistsics reported whether or not the means of the two components were significantly different or not. If they are significanty different they will describe two different gating events which might have significantly different physiological consequences.

Welcome to the baffling world of single channel analysis.