NOTE 02/06/2024 contents of this repository were transferred to: https://github.com/jacobkers/Cees_Dekker_Lab_Kerssemakers_Public_Code
Code associated with the paper:
Code associated with the paper: "Bulk-surface coupling reconciles Min-protein pattern formation in vitro and in vivo"
Fridtjof Brauns1,, Grzegorz Pawlik2,, Jacob Halatek3,, Jacob Kerssemakers2, Erwin Frey1,†, Cees Dekker2,† 1Arnold Sommerfeld Center for Theoretical Physics and Center for NanoScience, Department of Physics, Ludwig-Maximilians-Universität München, Theresienstraße 37, D-80333 München, Germany 2Department of Bionanoscience, Kavli Institute of Nanoscience Delft, Delft University of Technology, Van der Maasweg 9, 2629 HZ Delft, the Netherlands 3Biological Computation Group, Microsoft Research, Cambridge CB1 2FB, UKF.B., G.P., and J.H. contributed equally to this work. †Corresponding authors. Email: frey@lmu.de, c.dekker@tudelft.nl
Code was developed in the Cees Dekker lab by Jacob Kerssemakers
Version : November_11_2020_17_58
https://zenodo.org/badge/latestdoi/312041417
Order of listing: main programs: project-specific shell programs secondary programs are project-specific sub-programs tools are small sub-functions
Function list:
main analysis
A001_CleanSavePatterns A002_ResliceStacks A003_Reslice_and_CondenseStacks A400_Correlation A401_Correlation_Plotter A510_Top_and_Bottom_Correlation_XT
sub analysis
tools
Count_wavelength GetBackgroundImage JKD1_PRF_outlier_flag MedSmooth Scroll_ImageDirs Split_Image TrackXY_by_QI_Init
favorite tools
Full description:
main analysis
Name: A001_CleanSavePatterns Parameters in: no input Parameters out: no output main analysis Title: cleaning of images Summary: this function applies background cleaning of patterns. Background correction and artefact removal was done as follows: first , for movies, frames were corrected for fluorescence bleaching by normalizing each frame on its mean intensity value. This corrected for the max. 20 intensity decay over long movies. Next, two correction images were processed:
- a ‘static background’-image Imstat was made by averaging out all moving (wave pattern) features of the movie stack and removing any residual background level. Thus, this image only contained static fluorescent features such as specks, holes and scratches.
- an ‘illumination correction’ image Imillum was made by strongly
smoothening out and averaging all movie images and normalizing the result
to its maximum. Finally, each movie image Immovie was corrected as via the
following image operation: Imcorrected=(Immovie- Imstat)/Imillum .
This way, irregularities are suppressed and wave amplitudes on the edge of each image would not be underestimated compared to the amplitudes in the center of the image. Input: Paths are set at line ~~5. Movies should be stacked tif files Output: cleaned tif stacks, extension 'cln' Reference: Cees Dekker Lab, Project: MinED; researcher Grzegorz Pawlik; code designed & written Jacob Kerssemakers 2016
sourcename =
'A001_CleanSavePatterns'
Source code:A001_CleanSavePatterns line:2
Name: A002_ResliceStacks Parameters in: no input Parameters out: no output main analysis Title: reslicing of cleaned images Summary: this function tilts patterns: for quicker analysis, xy-t stacks are tilted x-t and y-t, such that each saved frame is essentially a kymograph Input: directory with pre-cleaned movies Output: per movie, an xt and a yt movie. Reference: Cees Dekker Lab, Project: MinED; researcher Grzegorz Pawlik; code designed & written Jacob Kerssemakers 2016
sourcename =
'A002_ResliceStacks'
Source code:A002_ResliceStacks line:2
Name: A003_Reslice_and_CondenseStacks Parameters in: no input Parameters out: no output main analysis Title: cleaning of images Summary: this function tilts patterns so that each frame represents a X-T kymograph Next, these are condensed in XY squares so the effect is, that one has a collection of time-boxes. These can be used for top-bottom correlation work (such as done with A510_Top_and_Bottom_Correlation_XT) Input: Output: Reference: Project: BN_CD16_Greg ; Jacob Kers 2016; Github
sourcename =
'A003_Reslice_and_CondenseStacks'
Source code:A003_Reslice_and_CondenseStacks line:2
Name: A400_Correlation
Parameters in: no input
Parameters out: no output
main analysis
Title: spatial and temporal correlation analysis
Summary: Spatial autocorrelation analysis was performed on 10 individual images per movie.
For each autocorrelation output image, a radial average was recorded
starting from the main central correlation peak.
The resulting spatial radial correlation curve was subjected to
maxima analysis. The first maximum after radius R=0 indicated
the most predominant distance between wave edges,
irrespective of propagation direction.
This distance was denoted as lambda. For temporal correlation,
we generated 20 x-t or y-t kymographs per movie
(10 in ‘x’ direction and 10 in ‘y’ direction)
evenly distributed over the middle 0.7 fraction of an image.
For each such kymograph, an autocorrelation analysis was performed.
The x=0 or y=0 line of these x-t autocorrelation maps then in effect
represents a temporal correlation curve averaged over all the original
image points on this line. Next, these correlation curves were
median averaged between different kymographs.
Thus, the final correlation curve in effect represents the average
temporal correlation signal sampled from 20x512 surface locations.
Analogous to the spatial correlation analysis, the first maximum after t=0
indicated a main oscillation period.
Experimental repeats of the same conditions (concentrations and height)
were median averaged.
Input: directory with pre-cleaned movies and directory with tilted xt or yt
versions of these movies, named following experimental
conditions. as follows:
First letter of the filename means height, ABCDEF from lowest to highest.
translates as heights: [2 6 8 15 25 57] microns
second comes number which will be E/D ratio 1,2,3,4,5 ,
translates as ratios: (0.5, 0.75. 1, 2, 3)
Last letter is channel (D -MinD, E -MinE)
Second last number is experiment session
Last number is repeat of experiment session with same conditions.
Example: 'C_3_E_1_2' means: height 8, ratio 1,MinE signal, session 1, repeat 2. Output:
- C_3_E_1_2_cln_overview_Height08Ratio2.00ChanERepeat01.jpg: plots summarizing the results of the spatial and temporal periodicity analysis
- A400_Correlation_results_permovie.mat: Data output description of Greg's 'correlation' data General ----------------------------------------- MinED_mov = 1×312 struct array with fields: filname: name of movie channel: colorlabel channel ('D' or 'E') height: sampe height EDratio: MinE-MinD ratio repeat: experimental repeat index of this EDratio-height combination xy: see below xt: see below yt: see below info: frames2mins: 1 pix2nm: 594
Details ----------------------------------------------
For example, MinED_mov(1).xy contains data of the first movie.
Each field xy or xt contains, per movie, some details of the obained correlation curves.
The xy correlation curves are obtained from radial averages of the
correlation maps, the xt correlation curves are taken along the x=0 line
(since then, we look at the autocorrelation of a wave with itself)
suffix _mv means per movie frame (here, 10 were chosen)
suffix _av means average of these values
data:
decay_val_labda_av:
decaylength_abs_av:
decaylength_sign_av:
decayperiodicityfactor_av:
envelope_abs: empty
example: [1×1 struct]:
mn1pos_av: position of first minimum (''anticorrelation'')
mn1val_av: amplitude of this peak
pk2pos_av: position of first maximum (indicative of periodicity)
pk2val_av: amplitude of this peak
r_axis: for plotting: axis values
radialprofiles: for plotting: profiles
strengthofmodulation_av: factor relating the modulation depth of the
correlation peaks to the overall decay value, to express how
'periodic' the waves look
- A400_Correlation_results_tabulated.mat: here, the above parameters are saved as columns, one row per movie
- A400_Correlation_results_tabulated.xlsx: as the .mat file
- A400_Correlation_results_matrices..jpg: visual representation of parameters, set in a matrix of heights and MinE/MinD ratio. Reference: Cees Dekker Lab, Project: MinED; researcher Grzegorz Pawlik; code designed & written Jacob Kerssemakers 2016
sourcename =
'A400_Correlation'
Source code:A400_Correlation line:2
Name: A401_Correlation_Plotter Parameters in: no input Parameters out: no output main analysis Title: matrix plotting of results Summary: This function perfoms matrix plotting of data processed by A400 Input: mat files in specified output dirspecified Output: matrix plots Reference: Cees Dekker Lab, Project: MinED; researcher Grzegorz Pawlik; code designed & written Jacob Kerssemakers 2016
sourcename =
'A401_Correlation_Plotter'
Source code:A401_Correlation_Plotter line:2
Name: A510_Top_and_Bottom_Correlation_XT Parameters in: no input Parameters out: no output main analysis Title: load 2 stacks and correlate them image by image; allows for timeshift Summary: This code is used to compare MinED patterns facing each other on the top and the bottom of a sample Input: movies taken at different sample heights are loaded. To ave time, these movies were tilted and resliced, so that xt-profiles can be loaded at once. Output: correlation data, in mat, excel and jpg Reference: Cees Dekker Lab, Project: MinED; researcher Grzegorz Pawlik; code designed & written Jacob Kerssemakers 2016
sourcename =
'A510_Top_and_Bottom_Correlation_XT'
Source code:A510_Top_and_Bottom_Correlation_XT line:2
sub analysis
tools
Name: Count_wavelength Parameters in: resliceim Parameters out: Lw tools Title: count wave peaks in image Input: kymograph Output: wavelength Reference: Cees Dekker Lab, Project: MinED; researcher Grzegorz Pawlik; code designed & written Jacob Kerssemakers 2019
sourcename =
'Count_wavelength'
Source code:Count_wavelength line:2
Name: Parameters in: addvals, diagnosefilename Parameters out: no output tools Title: compare different analysis routines in large analysis platforms Summary: This function collects two datasets; for example from two comparative analysis routines and saves the results in separate files. to avoid passing arrays or parameters through all functions, it initializes the container or just adds Input: data points or arrays, plus a nr for indicating options Output: mat files that are updated Reference: Cees Dekker Lab, Project: MinED; researcher Grzegorz Pawlik; code designed & written Jacob Kerssemakers 2019
sourcename =
1×0 empty char array
Source code: line:2
Name: GetBackgroundImage Parameters in: im,backsquaregridsize Parameters out: im_back tools Title: Get a background illumination image. Summary: Background Correction: take the medians of sub_regions or 'tiles' and smooth the result. The resulting image is taken as representative for an uneven illumination background Input: image, number of tiles per image width Output: background image, same size as original Reference: Cees Dekker Lab, Project: MinED; researcher Grzegorz Pawlik; code designed & written Jacob Kerssemakers 2016
sourcename =
'GetBackgroundImage'
Source code:GetBackgroundImage line:2
Name: JKD1_PRF_outlier_flag Parameters in: data,tolerance,sigchange,how,sho Parameters out: flag,cleandata tools this function is meant to find a representative value for a standard deviation in a heavily skewed distribution (typically, flat prf with peaks). It calculates the standard deviation and average the prf; Based on these, outliers are determined and excluded for a new calculation of average and SD; this is repeated until sigma does not change anymore too much This is repeated until the new sigma does not change much anymore output: positions of outliers. Jacob Kers 2013 and before input (suggested): data: single data array tolerance: beyond how many sigmas is considered outlier (3) sigchange: stop iteration if relative decrease of sigma is less (0.7) how: consider positive outliers, or all ('all' /'positive' sho: (for demo only) show intermediate graphs (0) output: flags: positions of outliers cleandata: data w/o outliers Reference: Cees Dekker Lab, code designed & written Jacob Kerssemakers 2016
sourcename =
'JKD1_PRF_outlier_flag'
Source code:JKD1_PRF_outlier_flag line:2
Name: MedSmooth Parameters in: data,window,how Parameters out: data_smz tools Title: median or average smoothing Input: array, options Output: smoothed array References: written by Jacob Kers, 2010 or so
sourcename =
'MedSmooth'
Source code:MedSmooth line:2
Name: Scroll_ImageDirs Parameters in: impth,SearchTemplate Parameters out: pic_list tools Title: Scroll image directories Summary:This function generates a list of full_pathfilenames and image names of sub-directories containing the right template string; to be used for expanded (image) data directories Input: source path, search template (*.[text]); Output: 1) structure list of pairs of full_pathfilenames and image names containing the right template string 2) list of directories that contain the specified template References: written by Jacob Kers, 2010 or so
sourcename =
'Scroll_ImageDirs'
Source code:Scroll_ImageDirs line:2
Name: Split_Image Parameters in: RawImage Parameters out: ObjectImage, IllumImage,ImProps tools Title: median or average smoothing Summary: Perform an iterative loop aimed at splitting the original image in two images: 1) one an 'illumination field' image, that represents a median-smoothend, 2D curved illumination field. 2) a flattened 'objects' image that only deviates from zero where surface objects are present; Input: image Output: two images, properties References: written by Jacob Kers, 2018 or so
sourcename =
'Split_Image'
Source code:Split_Image line:2
Name: TrackXY_by_QI_Init Parameters in: firstim Parameters out: pretracksettings tools Title: setup a radial sampling map Input: image Output: settings and sampling coordinates References: written by Jacob Kers, 2017 or so
sourcename =
'TrackXY_by_QI_Init'
Source code:TrackXY_by_QI_Init line:2