FRETBursts - 8-spot smFRET burst analysis

This notebook is part of a tutorial series for the FRETBursts burst analysis software.

For a step-by-step introduction to FRETBursts usage please refer to us-ALEX smFRET burst analysis.

In this notebook we present a typical FRETBursts workflow for multi-spot smFRET burst analysis. Briefly, we show how to perform background estimation, burst search, burst selection, FRET histograms, and FRET efficiency fit using different methods.

Loading the software

In [1]:
from fretbursts import *
 - Optimized (cython) burst search loaded.
 - Optimized (cython) photon counting loaded.
 You are running FRETBursts (version 0.6.4).

 If you use this software please cite the following paper:

   FRETBursts: An Open Source Toolkit for Analysis of Freely-Diffusing Single-Molecule FRET
   Ingargiola et al. (2016). 

In [2]:
sns = init_notebook()
In [3]:
import lmfit; lmfit.__version__
In [4]:
import phconvert; phconvert.__version__

Downloading the sample data file

The complete example dataset can be downloaded from here.

Here we download an 8-spot smFRET measurement file using the download_file in FRETBursts:

In [5]:
url = ''
In [6]:
download_file(url, save_dir='./data')
File: 12d_New_30p_320mW_steer_3.hdf5
File already on disk: /Users/anto/src/FRETBursts/notebooks/data/12d_New_30p_320mW_steer_3.hdf5 
Delete it to re-download.

Selecting a data file

In [7]:
filename = "./data/12d_New_30p_320mW_steer_3.hdf5"
In [8]:
import os
assert os.path.exists(filename)

Load and process the data:

In [9]:
d = loader.photon_hdf5(filename)

For convenience we can set the correction coefficients right away so that they will be used in the subsequent analysis. The correction coefficients are:

  • leakage or bleed-through: leakage
  • direct excitation: dir_ex (ALEX-only)
  • gamma-factor gamma

The direct excitation cannot be applied to non-ALEX (single-laser) smFRET measurements (like the current one).

In [10]:
d.leakage = 0.038
d.gamma = 0.43

NOTE: at any later moment, after burst search, a simple reassignment of these coefficient will update the burst data with the new correction values.

Compute background and burst search:

In [11]:
d.calc_bg(bg.exp_fit, time_s=30, tail_min_us='auto', F_bg=1.7)
d.burst_search(L=10, m=10, F=7)
 - Calculating BG rates ... [DONE]
 - Performing burst search (verbose=False) ...[DONE]
 - Calculating burst periods ...[DONE]
 - Counting D and A ph and calculating FRET ... 
   - Applying background correction.
   - Applying leakage correction.
   [DONE Counting D/A]

Perform a background plot as a function of the channel:

In [12]:

Let's take a look at the photon waiting times histograms and at the fitted background rates:

In [13]:
dplot(d, hist_bg);

Using dplot exactly in the same way as for the single-spot data has now generated 8 subplots, one for each channel.

Let's plot a timetrace for the background to see is there are significant variations during the measurement:

In [14]:
dplot(d, timetrace_bg);

We can look at the timetrace of the photon stream (binning):

In [15]:
dplot(d, timetrace)
xlim(2, 3); ylim(-100, 100);

We can also open the same plot in an interactive window that allows scrolling (uncomment the following lines):

In [16]:
#%matplotlib qt
In [17]:
#dplot(d, timetrace, scroll=True);
In [18]:
#ylim(-100, 100)
In [19]:
#%matplotlib inline

Burst selection and FRET

Selecting bursts by burst size (select_bursts.size)

In [20]:
gamma = d.gamma
In [21]:
d.gamma = 1
ds = d.select_bursts(select_bursts.size, th1=30, gamma=1)
dplot(ds, hist_fret);
In [22]:
ds = d.select_bursts(select_bursts.size, th1=25, gamma=gamma, donor_ref=False)
dplot(ds, hist_fret);
In [23]:
ds = d.select_bursts(select_bursts.size, th1=25, gamma=gamma)
dplot(ds, hist_fret, weights='size', gamma=gamma);
In [24]:
dplot(ds, scatter_fret_nd_na); ylim(0,200);

FRET Fitting

2-Gaussian mixture

Let's fit the $E$ histogram with a 2-Gaussians model:

In [25]:
ds.gamma = 1.
bext.bursts_fitter(ds, weights=None)
ds.E_fitter.fit_histogram(mfit.factory_two_gaussians(), verbose=False)

The fitted parameters are stored in a pandas DataFrame:

In [26]:
p1_amplitude p1_center p1_sigma p2_amplitude p2_center p2_sigma
0 0.161037 0.0147363 0.054972 0.827205 0.587919 0.144063
1 0.154906 0.0165125 0.065014 0.830402 0.592552 0.147132
2 0.172021 0.00363337 0.0429549 0.818251 0.56242 0.148298
3 0.151185 0.00576416 0.0410726 0.827219 0.565177 0.14754
4 0.155672 0.00718107 0.0442892 0.830064 0.558945 0.15308
5 0.169139 0.00913894 0.0535911 0.822227 0.561622 0.153229
6 0.15287 0.00624211 0.0480221 0.82579 0.559307 0.143679
7 0.154317 0.00772731 0.0412279 0.822823 0.556762 0.138972
In [27]:
dplot(ds, hist_fret, weights=None, show_model=True,
      show_fit_stats=True, fit_from='p2_center');