This version allows to fix diffusion coefficients in the model (in case those parameters have already been estimated from an independent source.
To use it, simply set the lower bound and the upper bound of the corresponding diffusion coefficient to the independently estimated value.
Implementation details: Look at the difference between LB and UB, if under a float threshold, consider that this parameter should be fixed, and declare it as is.
API change: this should allow not to break the API, and provide natural evolution for aplications that depend on fastspt, such as Spot-On.
- Started a Python port of Anders' code
- Including the plotting system
- Started a port in Python
Performed control experiments to measure the axial detection slice that can actually be determined using fixed cells labeled with JF549 or JF646 and z-stack imaging with 20 nm z-stack. It is difficult to say with absolue certainty, but dZ appears to be around 700 nm. Performed new Monte Carlo simulations and determined a new empirical relationship assuming dT = 4.477 ms and 1 gap allowed:
dZ_corr = dZ + a*sqrt(D) + b
dZ_corr = 700 nm + 0.15716*sqrt(D) + 208.11 nm
Updated the code to reflect this. SS_2State_model_Z_corr_v4 reflect
this update.
Re-wrote data loading section to allow for loading in data from multiple different experiments and fitting a curve to all of the data. This is mainly for figure visualization purposes, so do not use Single-Cell fitting when doing this.
Performed Monte Carlo simulations to find the optimal Z-corr for your
conditions: dZ = 0.8 um; dT = 4.5 ms; 1 gap allowed;
So now using the new empirical relationships:
dZ_corr = dZ + 0.144*sqrt(D) + 0.271;
Modify the code to allow cell-by-cell analysis. This will help identifying any biasing aspects of any particular movies and movies where the particle density was too high or too low can be excluded.
Since you are not accounting for transitions between bound and unbound
within DeltaTau because simulations showed this was negligble for CTCF,
there is no need to use
Correct for molecules that are moving in the axial dimension. The moving fraction is more likely to move out of focus. Use the procedure of Kues and Kubitscheck assuming an absorbing boundary and then use the empirical relationship from Davide Mazza:
DeltaZ_USE = DeltaZ_TRUE + 0.21 * sqrt(D); % (v4 uses different expression)
Also, you can now do both PDF and CDF fitting. The variable PDF_or_CDF
determines whether you are fitting to a histogram or to a CDF function:
-
PDF_or_CDF = 1==> perform histogram fitting -
PDF_or_CDF = 2==> perform CDF fittingThis function now includes a correction for the fact that free molecules are more likely to be lost and move out of focus.