I am working on Ionotropic glutamate receptors. Here I attached 100 ms, desensitized currents of GluR 6 Native with 10 mM Glutamate (abf) file. Please help to fit for desensitization ...
It looks pretty clearly like you've got multiple processes going on here. Are you sure you're applying a constant stream of glutamate, because it looks like something happens to the concentration of glutamate at about 66ms. A constant stream of glutamate should produce a pretty nice exponential process (see A Molecular Determinant for Submillisecond Desensitization in Glutamate Receptors). Also, I hope you're performing two electrode voltage clamp, or there is no way your voltage clamp is holding. However, ignoring those things, if you fit an exponential to the late phase of the decay (from 70 to 90ms) you get a moderately good fit, with a tau of 3.3 ms
Thanks for your help, I have some further questions regarding fitting of desensitization kinetics. It is not two electrode voltage clamp recording. I performed this experiment with ultrast theta pipette based perfusion system with normal voltage clamp recording. Holding at -60 mV, the recording was performed with HEKA amplifier and traces were converted in to abf file.
If possible can you share the screen shots or result file.
1. All my traces are fitting with two-term exponential fits. So some of the traces the Tau 1 and Tau 2 values are having significant variations. In that case which one I have to take for final consideration (either Tau1 or Tau2)?
2. Among all the fitting methods ( levenberg-Marquardt Method, Variable metric method, Simplex, and Chebyshev methods) Chebyshev methods are giving a lot of variation with some of my recordings and it is a default method pClamp. Usually, for voltage-gated channels, Tau 1 is used to calculate the fast inactivating currents and Tau 2 is for slow inactivating currents. I have no clue regarding the ultra-fast piezo-based ligand application evoked currents ad I have gone through several papers but no method is discussed in detailed fitting kinetics for this channel currents
3. Exponential fitting: I am using the exponential fitting to the decay of current from ~90 % of its peak (cursor no.1) amplitude to baseline and having a time constant of τdeact (cursor no 2) for whole cell macroscopic rate of deactivation. is the way to place the cursors?
Well perhaps there was an error in the conversion from HEKA data to ABF. I hope there was. Because you can't pass nearly 15 nA if you're doing single electrode voltage clamp and expect to have voltage clamp. Multiply that current by your series resistance and figure out how large your voltage clamp error is? 30mV at a minimum I would guess. I'm not sure what kind of cells you are in, but perhaps the voltage clamp error could explain why your waveform looks the way it does.
I believe you have a problem with your application. Applying a constant concentration of glutamate should give you , roughly speaking, a monoexponential decay, not the complex waveform you see, at least in my experience and as one sees in published papers. Because of this, any attempt to fit a mono, or biexponential from the 90% to 0% decay of this curve is futile, as you do not have a mono or biexponential. You might as well try to fit a sin wave and talk about the frequency of the decay.
Your point 1. I'm not sure what your question is. And I'm not sure what kind of variation you are talking about. If you have a biexponential process, and you fit a biexpoential to it, you would generally expect tau1 and tau2 to be quite different. You do not take one of the values and call that your result. You report both of them. Some people report a "weighted" decay constant, which is the average of the taus, weighted by their respective amplitudes.
Your point 2. Again, I'm not sure what variation you are talking about. Different recordings will give you different answers. What is important is that the fit and the results make sense.
Your point 3. As I've said, you cant fit an exponential to non exponential data and expect it to make sense. Hence fitting an exponential to the 90%-0% data is pointless.
That said, you do appear to have something close to a monoexponential in the late phase of the decay