BioPhyPy: A python library for the analysis of biophysical data

This library contains modules for the analysis of different biophysical techniques. Most of the them are experimental. No warranty whatsoever. DSF_fit, MST_data and MP_data are available as user-friendly webservers: https://spc.embl-hamburg.de/

Please cite the respective publications and acknowledge this repository and the SPC facility at EMBL Hamburg when using it for a publication:

• DSF_fit
  • Niebling, S., Burastero, O., Bürgi, J., Günther, C., Defelipe, L. A., Sander, S., ... & García-Alai, M. (2021). FoldAffinity: binding affinities from nDSF experiments. Scientific reports, 11(1), 1-17.
  • Burastero, O., Niebling, S., Defelipe, L. A., Günther, C., Struve, A., & Garcia Alai, M. M. (2021). eSPC: an online data-analysis platform for molecular biophysics. Acta Crystallographica Section D: Structural Biology, 77(10).
  • Bai, N., Roder, H., Dickson, A., & Karanicolas, J. (2019). Isothermal analysis of ThermoFluor data can readily provide quantitative binding affinities. Scientific reports, 9(1), 1-15.
• MP_data
  • Niebling, S., Veith, K., Vollmer, B., Lizarrondo, J., Burastero, O., Schiller, J., ... & García-Alai, M. (2022). Biophysical Screening Pipeline for Cryo-EM Grid Preparation of Membrane Proteins. Frontiers in Molecular Biosciences, 535.

Installation

There is no installation necessary. After cloning the repository, add the parental folder to the path and then load the module you want to use. Below is an example:

import sys
# Load parental folder that contains BioPhysPy folder
sys.path.append("../")
# Load python class
from BioPhyPy.DSF_fit import DSF_binding

Modules

Currently BioPhyPy contains the following modules:

• DSF_fit: Differential scanning fluorimetry
• MP_data: Mass photometry
• BLI_data: Biolayer interferometry
• MST_data: Microscale thermophoresis

DSF_fit: Differential scanning fluorimetry

• Visualization of DSF data
• Isothermal analysis
• Tm analysis

Load and plot DSF data

# Load modules
import sys
# Add parental folder that contains module to path
sys.path.append("../")
# Load module
from BioPhyPy.DSF_fit import DSF_binding

# Parameters, adjust!
fn = './2405_ndsf_example/example.xlsx'
# Decide whether to load '330nm', '350nm' or 'Ratio' data
which = 'Ratio'

# Load data
test = DSF_binding()
# Specify temperature window
window = [22, 55]

# Create list with analyte concentrations
# Here it was a serial dilution with 14 concentrations
# starting at 5 mM
concs = []
for i in range(14):
    concs.append(5E-3*0.5**i)
# Pure protein without analyte
concs.append(0)
# Experiment was run in triplicates
concs = 3*concs

# Load data
test.load_xlsx(fn=fn, concs=concs, which=which, window=window, load_fit=False)
# Protein Concentration
test.pconc = 10E-6
# Plot fluorescence signal
fig, ax = test.plot_fluo()
# Save figure
fig.savefig('./fluo.pdf')

Do isothermal analysis to obtain Kd

# Load modules
import sys
# Add parental folder that contains BioPhysPy to path
sys.path.append("../")
# Load python class
from BioPhyPy.DSF_fit import DSF_binding

# Parameters, adjust!
fn = './2405_ndsf_example/example.xlsx'
# Decide whether to load '330nm', '350nm' or 'Ratio' data
which = 'Ratio'

# Load data
test = DSF_binding()
# Specify temperature window
window = [22, 55]

# Create list with analyte concentrations
# Here it was a serial dilution with 14 concentrations
# starting at 5 mM
concs = []
for i in range(14):
    concs.append(5E-3*0.5**i)
# Pure protein without analyte
concs.append(0)
# Experiment was run in triplicates
concs = 3*concs

# Load data
test.load_xlsx(fn=fn, concs=concs, which=which, window=window, load_fit=False)
# Protein Concentration
test.pconc = 10E-6

# Local fit of fluorescence curves
test.fit_fluo_local()
# Subsequent global fit
test.fit_fluo_global()
# Plot fits
fig, ax = test.plot_fit_fluo()
fig.savefig('./fluo_fit.pdf')
# Define temperatures for isothermal analysis, should be around Tm
test.isothermal_ts = [32,34,36,38, 40, 42, 44, 46, 48] 
# Isothermal analysis
test.fit_isothermal()
# Plot isothermal analysis
fig, ax = test.plot_fit_isothermal(show_only_kd=True)
fig.savefig('./iso.pdf')

Melting temperature analysis

# Load modules
import sys, os
# Add parental folder that contains BioPhysPy to path
sys.path.append("../")
# Load python class
from BioPhyPy.DSF_fit import DSF_binding

# Parameters, adjust!
fn = './2405_ndsf_example/example.xlsx'
# Decide whether to load '330nm', '350nm' or 'Ratio' data
which = 'Ratio'

# Load data
test = DSF_binding()
# Specify temperature window
window = [22, 55]

# Create list with analyte concentrations
# Here it was a serial dilution with 14 concentrations
# starting at 5 mM
concs = []
for i in range(14):
    concs.append(5E-3*0.5**i)
# Pure protein without analyte
concs.append(0)
# Experiment was run in triplicates
concs = 3*concs

# Load data
test.load_xlsx(fn=fn, concs=concs, which=which, window=window, load_fit=False)
# Protein Concentration
test.pconc = 10E-6

# Determine melting temperatures from 1st derivative 
test.tms_from_derivatives()
# Specify fitting model: 'single' site or 'hill'
tms_fit = 'single' # 
test.fit_tms(fit=tms_fit)
# Plot melting temperatures and fit
fig, ax = test.plot_tms()
fig.savefig('./tms_%s.pdf' % tms_fit)

MP_data: Mass photometry

• Visualization of mass photometry data
• Generation of histograms and gaussian fitting
• Can also read movie file to show frames in combination with histogram

Mass photometry histogram

import sys
import matplotlib.pyplot as plt
# Add parental folder that contains BioPhysPy to path
sys.path.append("../")
# Load python class
from BioPhyPy import MP_data

# Adjust font size
fs = 18
plt.rcParams.update({'font.size': fs})

# Histogram file
# This is the linux path syntax, for Windows it needs to be adjusted, e.g. by using \ instead of /
fn ='./folder/example.h5'

# Default parameters for processing
# Limits for
show_lim = [-100, 600]
guess_pos = [66, 148, 480]
max_width=20
# Load eventsFitted
dataset = MP_data(fn=fn, mp_fn=mp_fn)
# Create histogram
dataset.create_histo(bin_width=4)
# Fit bands in mass space
dataset.fit_histo(guess_pos=guess_pos, tol=30, cutoff=0, xlim=[-100,1500], max_width=max_width)
fig.savefig('./mst_initfluo.pdf')
# Create wider plot than default
fig, ax = plt.subplots(1, figsize=[8, 4])
# Plot histogram
dataset.plot_histo(xlim=show_lim, ax=ax, show_labels=True, short_labels=True, show_counts=True)

# Save plot
fig.savefig('./mp_plot.pdf')

Optional: Showing frame from movie as inlet

import sys, glob, os
# Add parental folder that contains BioPhysPy to path
sys.path.append("../")
import numpy as np
# Load python class
from BioPhyPy import MP_data

# Font size
fs = 18
plt.rcParams.update({'font.size': fs})

# Histogram file
# This is the linux path syntax, for Windows it needs to be adjusted, e.g. by using \ instead of /
fn ='./folder/events.h5'
# Movie file
mp_fn = './folder/movie.mpr'

# Default parameters for processing
# Limits for
show_lim = [-100, 600]
guess_pos = [66, 148, 480]
max_width=20
# Load eventsFitted and movie file
dataset = MP_data(fn=fn, mp_fn=mp_fn)
# Set parameters for later plot of the frame
# Analyze movie for later inlet
# Frame with largest numbers of fitted events above threshold is chosen
dataset.analyze_movie(frame_range=2, frame='largest', threshold=50, show_lines=True)
# Create histogram
dataset.create_histo(bin_width=4)
# Fit bands in mass space
dataset.fit_histo(guess_pos=guess_pos, tol=30, cutoff=0, xlim=[-100,1500], max_width=max_width)

# Create plot
fig, ax = plt.subplots(1, figsize=[8, 4])
dataset.plot_histo(xlim=show_lim, ax=ax, show_labels=False, short_labels=True, show_counts=False, contrasts=False)

# Save plot
fig.savefig('./mp_inlet.pdf')

BLI_data: Bilayer interferometry

• Visualization of BLI data
• Kinetic fits not fully implemented yet

Showing sensograms

# Load modules
import sys
# Add parental folder that contains BioPhysPy to path
sys.path.append('../')
# Load module
from BioPhyPy import BLI_data

# Font size
fs = 12
matplotlib.rcParams.update({'font.size': fs})

# Folder with raw data
# This is the linux path syntax, for Windows it needs to be adjusted, e.g. by using \ instead of /
folder = './experiment_folder/'

# Initialize instance
bli_data = BLI_data(folder=folder)
# Remove jumps (use average of first 3 points)
bli_data.remove_jumps(xshift=3)
# Align curves to beginning of association (step 3)
bli_data.align(step=0, location='start')
# Smooth curves with a 21 point window
bli_data.smooth(window_length=21)

# Plot signal for binding
fig, ax = bli_data.plot(legend='SampleID', legend_step=3, abbrev_step_names=True, steps=[0,1,2,3,4], sensors=range(1,8))
# Save plot### MST_data
* Visualize data and fit affinities
fig.savefig('./bli_plot.pdf')

Example for kinetic fit (experimental)

# Load modules
import sys
# Add parental folder that contains BioPhysPy to path
sys.path.append('../')
from BioPhyPy import BLI_data

# Font size
fs = 12
matplotlib.rcParams.update({'font.size': fs})

# Folder with raw data
# This is the linux path syntax, for Windows it needs to be adjusted, e.g. by using \ instead of /
folder = './experiment_folder/'

# Initialize instance
bli_data = BLI_data(folder=folder)
# Remove jumps (use average of first 3 points)
bli_data.remove_jumps(xshift=3)
# Align curves to beginning of association (step 3)
bli_data.align(step=2, location='end')
# Smooth curves with a 21 point window
bli_data.smooth(window_length=21)
# # Subtract reference sensor
bli_data.subtract(ref_sensor=7, sample_sensors=[0,1,2,3,4,5,6,7])
# Align again
bli_data.align(step=4, location='end')

# Do fit of selected sensor and plot it
fig, ax = bli_data.fit_data(sensors=[4], step_assoc=3, step_dissoc=4, func='monoexp', plot=True)
# Save plot
fig.savefig('./bli_fit.pdf')

MST_data: Microscale thermophoresis

• Visualize data and fit affinities

Do MST analysis

# Load modules
import sys
import matplotlib.pyplot as plt
# Add parental folder that contains BioPhysPy to path
sys.path.append('../')
# Load module
from BioPhyPy import MST_data

# Folder
fn = './mst_example.xlsx'

# Define fontsize for plots
fs = 14
plt.rcParams.update({'font.size': fs})

# Load data
data = MST_data(fn=fn)
# Normalize signal
data.normalize()
# Select "hot" region for MST analysis (seconds)
data.calc_fnorm(hot=20)
# Define protein concentration
data.pconc = 10E-6
# Fit Kd with fixed protein concentration
data.get_kd(fix_pconc=True)
# Plot data
fig, ax = data.plot()
# Save figure 
fig.savefig('./mst.pdf')

Check initial fluorescence

# Load modules
import sys
import matplotlib.pyplot as plt
# Add parental folder that contains BioPhysPy to path
sys.path.append('../')
# Load module
from BioPhyPy import MST_data

# Folder
fn = './2405_mst_example/mst_example.xlsx'

# Define fontsize for plots
fs = 14
plt.rcParams.update({'font.size': fs})

# Load data
data = MST_data(fn=fn)
# Select "hot" region for MST analysis (seconds)
data.calc_fnorm(hot=20)

# Plot data
fig, ax = data.plot_init_fluo()

Discontinued modules

• CD_data
• IR_data
• MS_data
• DLS_data