examples: determine effective kon #713
Conversation
aafkevandenberg
force-pushed
the
example_kon
branch
from
44566e7 to
2bf5667
Compare
aafkevandenberg
requested review from
tommasogritti,
tobiasjj and
JoepVanlier
|
|
||
| Fit an exponential ditribution to the distribution of time intervals using Pylake:: | ||
|
|
||
| single_exponential_fit = lk.DwelltimeModel(np.array(intervals_total), n_components=1) |
There was a problem hiding this comment.
I think it would make sense to give some thought on discretization and the minimally observable time.
Discretization is easy, that's just the line time. The latter I would expect to be related to the gap window size in some way since you can't really observe on times shorter than this (because they would be connected to the adjacent track).
For this data, it isn't really needed, given that the binding times you measure here are quite long (36 seconds). In that case, I don't expect it to change the answer much in this case (since your line time is much shorter), but I think it would make the notebook a bit more generally usable for people who have short on timescales.
What do you think?
| In this Notebook, we will determine the binding time of a fluorescently labeled protein binding to DNA. | ||
| The protein binds and unbinds to target sites on DNA and the result is recorded as a kymograph. |
There was a problem hiding this comment.
For this one, and also the other binding notebook actually, we should probably mention that we neglect photobleaching here (which would lengthen t_on and shorten t_off). Users should make sure that their samples aren't expected to bleach on the timescales that they are considering.
One paper that might be interesting to look at in the future would be this one, though it needs a very large number of counts (I wouldn't suggest that for this analysis right now, as I doubt this amount of data would work reliably there). There are more pragmatic ways to correct for bleaching too, but given that we haven't tested those, I would not explicitly recommend a particular one here.
|
|
||
| def time_intervals(tracks): | ||
| """Compute the time intervals between all tracks in a given selection""" | ||
| intervals = [tracks[x+1].seconds[0]-tracks[x].seconds[-1] for x in range(len(tracks)-1)] |
There was a problem hiding this comment.
You rely on tracks being sorted with respect to time and non-overlapping here. Unfortunately, tracks in a group are not guaranteed to be sorted in time. While this is usually the case for track_greedy (but only when you use a single ROI), it is not for track_lines for instance
I would sort them and throw if you encounter overlaps.
The second thing you rely on is that this is a KymoTrackGroup for a single kymo. Unfortunately, there is no public API to validate that right now (internally we use _validate_single_source for that when a method requires that it is not a KymoTrackGroup with multiple sources).
Co-authored-by: Joep Vanlier <[email protected]>
Co-authored-by: Joep Vanlier <[email protected]>
Co-authored-by: Joep Vanlier <[email protected]>
I added an example for a frequently requested workflow: kon.
Docs can be found here: https://lumicks-pylake.readthedocs.io/en/example_kon/examples/kon/kon.html