8. Image and Signal Analysis

Useful Tools/Papers

Experiment Design

A Statistical Framework for Robust and Reproducible BioImage Analysis

Calcium Image Analysis

suite2p: Calcium imaging analysis.
EZcalcium: Calcium imaging analysis. Motion correction, ROI detection, ROI refinement, and ROI matching.
PatchWarp: Calcium imaging analysis. Corrects complex image distortions that slowly occur during long imaging sessions.

Cell Segmentation

Cellpose: Cell and nucleus segmentation
ilastik: Leverage machine learning algorithms to easily segment, classify, track and count your cells or other experimental data.

Spots Detection

RS-FISH

Motion Tracking

Napari B-track

ImageJ / Fiji Macro

Save as tiff

Some image processing software only work with tiff files. This scripts reads microscope images in their original formats(e.g. czi, nd2, lsm) and saves them as tiff.

// saveastiff_batch.ijm
// Author: Jane Ling (yan_to.ling@kcl.ac.uk)
// Last modified: 2025-02-18 15:23
//
// This macro saves images in fileDir as tif in saveDir.
// Note that image file names cannot contain " ". Use "_" or "-" instead.
//
// INPUTS
// - microscope images (e.g. czi, nd2, lsm)
//
// OUTPUTS
// If not splitting channels, ...
// - (original filename)_raw.tif
// If splitting channels, ...
// - C1-(original filename)_raw.tif
// - C2-(original filename)_raw.tif
// - C3-(original filename)_raw.tif

// MODIFY HERE: Path of the image files to be processed, searches all sub-directories.
fileDir = "/PATH/TO/PROJECT/originals/";
saveDir = "/PATH/TO/PROJECT/outputs/";

// MODIFY HERE: end of filename and input file type (file extension) e.g.".lsm"
endString = ".lsm";

// MODIFY HERE: whether to split channels and save as separate tif
splitChannel = true; 

///////////////////////////////////////////////////

setBatchMode(true);

print("Saving tif...");

walkFilesSaveTif(fileDir, saveDir, splitChannel, endString);

print("Done:)");


////////////////////////////////////////////////////
//                Local Functions                 //
////////////////////////////////////////////////////

// Find all files in subdirs, split channels and save tiff
function walkFilesSaveTif(inputDir, outputDir, splitChannel, endString) {
	list = getFileList(inputDir);

	// create output folder it it does not exist
	if (!File.exists(outputDir)) {
		File.makeDirectory(outputDir);
		}
	
	for (i=0; i<list.length; i++) {
		// If folder
		if (endsWith(list[i], File.separator))
		    walkFilesSaveTif( inputDir+list[i], outputDir+list[i], splitChannel, endString);
		    
		// If image file
		else  if (endsWith(list[i], endString)) {
			print (list[i]);
			// for saving all outputs in the same folder
			// splitChannels(inputDir, outputDir, list[i]); 

			// for creating a separate folder for each image file
			outputDir1 = outputDir + File.getNameWithoutExtension(list[i]) + File.separator;
		    saveTif(inputDir, outputDir1, list[i], splitChannel); 
		}
	 }
} // end of walkFilesSaveTif

// Split channels and save as tiff
function saveTif(inputDir, outputDir, file, splitChannel) {
	open(inputDir + file);

	// create output folder it it does not exist
	if (!File.exists(outputDir)) {
		File.makeDirectory(outputDir);
	}
	
	if (splitChannel) {
		// total number of channels
		Stack.getDimensions(width, height, channels, slices, frames);
		Nc = channels;
		
		run("Split Channels");
		
		for (i=0; i<Nc; i++){
			title = getTitle();
			title = substring(title, 0, lastIndexOf(title, ".")); // get first part of file name
			title = title + "_raw";
			
			// if already saved, close 
			if (File.exists(outputDir + title + ".tif")) {
				close();
			}
			else { // otherwise, save as tif
				saveAs("Tiff", outputDir + title); 
				close();
			}
		}
	}
	else {
		title = getTitle();
		title = substring(title, 0, lastIndexOf(title, ".")); // get first part of file name
		title = title + "_raw";
		
		// if already saved, close 
		if (File.exists(outputDir + title + ".tif")) {
			close();
		}
		else { // otherwise, save as tif
			saveAs("Tiff", outputDir + title); 
			close();
		}
	}
} // end of saveTif

Export ROI as CSV

This script saves ROI boundaries as a list of coordinates in csv, which could be loaded later for processing.

// export_ROIs_as_csv.ijm
// Author: Jane Ling (yan_to.ling@kcl.ac.uk)
// Last modified: 2025-02-18 15:23
//
// This macro saves multiple ROIs from ROI Manager as csv
//
// Instructions
// 1. Open the image
// 2. Use any of the selection or line tool to draw an ROI.
// 3. Go to "Edit" -> "Selection" -> "Add to Manager". 
// 4. Check that in "More >>" -> "Options", the checkbox "Associate "Show All" ROIs with slices" is checked.
// 5. Draw all other ROIs in the image/stack and add them to ROI Manager.
// 6. Go to"File" -> "Open".
// 7. Open the ImageJ macro file export_ROIs_as_csv.ijm
// 8. Modify fileDir to the saving directory on your computer.
// 9. Click Run to run the script. (Do NOT close the image before running the script.)
//
// INPUTS
// - ROIs from ROI Manager
//
// OUTPUTS
// - XY_(original filename)_1.csv
// - XY_(original filename)_2.csv
// - ...
// - (original filename)-ROIset.zip


print("Saving ROIs...");

// MODIFY HERE: Path to saving directory
fileDir = "/PATH/TO/PROJECT/outputs/";

// MODIFY HERE: Whether to create new folder under the saving directory with the name of the image
newFolder = true;

///////////////////////////////////////////////////

title = getTitle();
title = substring(title, 0, lastIndexOf(title, ".")); // get first part of file name

if (newFolder) {
	fileDir = fileDir + title + File.separator;
	// create output folder it it does not exist
	if (!File.exists(fileDir)) {
		File.makeDirectory(fileDir);
	}
}

for (i=0; i<RoiManager.size; i++) {
	roiManager("Select", i);
	
	results_csv_path = fileDir + "XY_" + title + "_" + (i+1) + ".csv";
	print(results_csv_path);
	
	Roi.getCoordinates(x, y);
	slice = getSliceNumber();
	Stack.getPosition(channel, slice, frame);

	run("Clear Results");
	
	for (k=0; k<x.length; k++) {
		setResult("X",k,x[k]);
		setResult("Y",k,y[k]);
		setResult("Z",k,slice);
	}
	updateResults();
	selectWindow("Results");
	saveAs("Results", results_csv_path);
	
	roiManager("save", fileDir + title + "-ROIset.zip");
}

print("Done:)");

Matlab

Here are some materials for those who want to use MATLAB for image analysis.

Useful MATLAB functions

Starting a Script

% close all figures
close all

% clear workspace
clear

% clear command window
clc

% Change the current folder to the folder of this m-file.
if(~isdeployed)
  cd(fileparts(matlab.desktop.editor.getActiveFilename));
end

Loading and Saving Data / Figures

The Bio-Formats Library

Bio-Formats is a convenient toolbox for reading biomedical images in various file formats. It has been installed on Clarke for Matlab under the folder /usr/local/bin/MATLAB/bfmatlab.

To load any data file, use the following lines:

% path to bioformat toolbox % change path if necessary
addpath('/PATH/TO/FOLDER/bfmatlab/')
ops.filename = "/PATH/TO/FOLDER/filename";
[im_data, ops] = loadBioFormats(ops);

%% [im_data, ops] = loadBioFormats(ops, varargin)
% Loads data using bioformats toolbox and reshapes data.
%
% Author: Jane Ling (yan_to.ling@kcl.ac.uk)
% Last updated: 2025-02-14 12:40
%
% USAGE:
% 1) [im_data, ops] = loadBioFormats(ops);
% 2) [im_data, ops] = loadBioFormats(ops, verbose);
%
% INPUTS:
%   - ops 
%       (struct) ops.filename defining the path to image file
%   - verbose 
%       (bool) whether to pring information about dataset or not.
%       (Default) false
%
% OUTPUTS:
%   - im_data
%       (numeric) 5D matrix if dimension order starts with 'XY', singular
%                 dimensions would be omitted.
%       (cell array) interleaved frames otherwise
%   - ops
%       (struct) parameters of the stack

function [im_data, ops] = loadBioFormats(ops, varargin)
    
    if nargin == 2
        verbose = varargin{1};
    else 
        verbose = 0;
    end

    % Bio-Formats Toolbox
    [~,~] = bfCheckJavaPath();   % added such that path to Bio-Formats Toolbox is know
    
    data = bfopen(ops.filename); % load data

    im_data = data{1,1};
    im_data = im_data(:,1); % image data
    
    omeMeta = data{1,4}; % metadata
    
    ops.imageIndex = 0;
    
    % size of stack in each dimension
    ops.Nx = omeMeta.getPixelsSizeX(ops.imageIndex).getValue(); % image width, pixels
    ops.Ny = omeMeta.getPixelsSizeY(ops.imageIndex).getValue(); % image height, pixels
    ops.Nz = omeMeta.getPixelsSizeZ(ops.imageIndex).getValue(); % number of Z slices
    ops.Nc = omeMeta.getPixelsSizeC(ops.imageIndex).getValue(); % number of channels
    ops.Nt = omeMeta.getPixelsSizeT(ops.imageIndex).getValue(); % number of time points
    
    % dimension order in data
    ops.dimOrder = char(omeMeta.getPixelsDimensionOrder(ops.imageIndex).getValue());
    
    % size of pixels (voxels)
    ops.sizeX = double(omeMeta.getPixelsPhysicalSizeX(ops.imageIndex).value(ome.units.UNITS.MICROMETER)); % [microns]
    ops.sizeY = double(omeMeta.getPixelsPhysicalSizeY(ops.imageIndex).value(ome.units.UNITS.MICROMETER)); % [microns]
    
    % axes
    ops.x = (0:ops.Nx-1)*ops.sizeX;
    ops.y = (0:ops.Ny-1)*ops.sizeY;
    
    if verbose
        fprintf('Size of image = %d x %d. \nNumber of Z slices = %d. \nNumber of channels = %d. \nNumber of time points = %d. \n', ops.Nx, ops.Ny, ops.Nz, ops.Nc, ops.Nt)
        fprintf('Dimension order = %s  \n', ops.dimOrder)
    end

    % for z-stack
    if ops.Nz > 1
        try
            ops.sizeZ = double(omeMeta.getPixelsPhysicalSizeZ(ops.imageIndex).value(ome.units.UNITS.MICROMETER)); % [microns]
            ops.z = (0:ops.Nz-1)*ops.sizeZ;
            if verbose
                fprintf('Voxel size = %.3f µm x %.3f µm x %.3f µm  \n', ops.sizeX, ops.sizeY, ops.sizeZ)
            end
        catch
            if verbose
                fprintf('Pixel size = %.3f µm x %.3f µm  \n', ops.sizeX, ops.sizeY)
            end
            warning('Could not read sizeZ from omeMeta.')
        end
    else
        if verbose
            fprintf('Pixel size = %.3f µm x %.3f µm  \n', ops.sizeX, ops.sizeY)
        end
    end
    
    % for time stack
    if ops.Nt > 1
        try
            ops.sizeT = omeMeta.getPixelsTimeIncrement(ops.imageIndex).value(ome.units.UNITS.SECOND);
            ops.t = (0:ops.Nt-1)*ops.sizeT;
            if verbose
                fprintf('Pixel time imcrement = %.3f s  \n', ops.sizeT)
            end
        catch
            warning('Dimension order may be incorrect.')
        end
    end
        
    % reshaping
    if strcmp(ops.dimOrder(1:2),'XY') 
        im_data = cell2mat(im_data);
        im_data = reshape(im_data,ops.Ny,ops.Nc*ops.Nz*ops.Nt,ops.Nx);
        im_data = permute(im_data, [1,3,2]);
        N = [ops.Ny,ops.Nx, 0, 0, 0];
    
        N(strfind(ops.dimOrder,'C')) = ops.Nc;
        N(strfind(ops.dimOrder,'Z')) = ops.Nz;
        N(strfind(ops.dimOrder,'T')) = ops.Nt;
        im_data = reshape(im_data,N(1),N(2),N(3),N(4),N(5));

        im_data = squeeze(im_data);
        ops.dimOrder = ops.dimOrder(N~=1);
        if verbose
            fprintf('Dimension order of im_data after reshaping = %s  \n', ops.dimOrder)
        end
    end
    
end

If you see and error that says Unrecognized function or variable, you may check if the folder has been added to path in HOME -> Set Path.

If you see an error related to java memory size, try to increase the java heap size following the steps describe in MATLAB Preferences (R2010a+)

For additional information, read Using Bio-Formats in MATLAB

Loading

Function	Description
bfopen	Loads data using Bio-Formats toolbox
loadBioFormats	Custom function to load bioformats
load	Load variables from file (.mat) into workspace
imread	Read image from graphics file
readtable	Create table from file. For loading .xlsx or .csv

Saving

Function	Description
writetable	Write table to file
save	Save variables from workspace to file
saveas	Save figure to specific file format
imwrite	Write image to graphics file
saveastiff	Save multipage TIFF stack

Creating File and Folder Names (string manipulation)

Function	Description
fullfile	Build full file name from parts
strsplit	Split string or character vector at specified delimiter
strcat	Concatenate strings horizontally
filesep	File separator for current platform
fileparts	Get parts of file name
sprintf	Format data into string or character vector

Data Types

MATLAB datatypes

Function	Description
double	Double-precision arrays
single	Single-precision arrays
int8	8-bit signed integer arrays
logical	Convert numeric values to logicals
string	String array
struct	Structure array
table	Table array with named variables that can contain different types

Reshaping / Interpolation

Function	Description
cell2mat	Convert cell array to ordinary array of the underlying data type
num2cell	Convert array to cell array with consistently sized cells
reshape	Reshape array by rearranging existing elements
permute	Permute array dimensions
repmat	Repeat copies of array
interp1	1-D data interpolation (table lookup). See also interp2, interp3, interpn
length	Length of largest array dimension
size	Array size
squeeze	Remove dimensions of length 1
vertcat	Concatenate arrays vertically
horzcat	Concatenate arrays horizontally

Spatial Filtering and Intensity Adjustment

Function	Description
medfilt2	2-D median filtering. See also medfilt3
imgaussfilt	2-D Gaussian filtering of images. See also imgaussfilt3
fspecial	Create predefined 2-D filter
imfilter	N-D filtering of multidimensional images
imadjustn	Adjust intensity values in N-D volumetric image

Temporal Filtering

Function	Description
filter	1-D digital filter
movmean	Moving mean
movmedian	Moving median

Zero-padding

Zero padding helps to eliminate boundary/border effects of filtering by filling in the off-the-edge image pixels. Here is an example of zero padding by mirror-reflecting the array across the array border. For other common padding options, you may refer to the builtin imfilter function.

%% zero padding: mirror-reflected boundary
function data = zero_padding(data, zeropad_len)
    Z1 = flipud(data(1:zeropad_len, :));
    Z2 = flipud(data(end-zeropad_len+1:end,:));
    data = [Z1; data; Z2];
end

function data = remv_zero_padding(data, zeropad_len)
    data = data(zeropad_len+1:length(data)-zeropad_len,:);
end

Baseline Detection

%% Sliding window filter
function baseline_pc_median = sliding_window_filter(data, baseline_percentage, window)
    tic;
    disp('Applying sliding window filter...')
    zeropad_len = window;
    data = zero_padding(data, zeropad_len);
    T = size(data,1);
    baseline_pc_median = zeros(size(data));
    for k=1:T %for all timepoints
        kymo_sample = data(max(1,k-window/2):min(T,k+window/2),:); %take window around each timepoint
        sortedVals = sort(kymo_sample); %sort in ascending order
        baseline_sorted_percentage=sortedVals(1:round(baseline_percentage*size(kymo_sample,1)),:); %take a specified percentage of these values
        baseline_pc_median(k,:) = median(baseline_sorted_percentage); %find median of this percentage
    end
    baseline_pc_median = remv_zero_padding(baseline_pc_median, zeropad_len);
    toc
end

Edge Detection

%% Gaussian derivative filter
function edge = Gaussian_Derivative_Filter_padded(data, thres)
    tic;
    disp('Applying 1st order Gaussian filter...')

    % sigma for guassian filter
    sigma_y = 2;
    
    % length of zero-padding
    zeropad_len = 50;
    
    % zero-padding
    edge = zero_padding(data, zeropad_len);
    
    % 1st Order Derivative 1D Gaussian filter, detects slopes in time dimension
    edge = Gaussian_Derivative_Filter(edge, sigma_y);
    
    % remove zero-padding
    edge = remv_zero_padding(edge, zeropad_len);
    
    % correct for direction of the slope
    edge = -edge;
    
    % thresholding to remove slow flucturations in signal
    edge(edge<thres)=0;

    toc
end

function im = Gaussian_Derivative_Filter(im, sigma_y)
    y_values = -ceil(4*sigma_y):ceil(4*sigma_y);
    filter_y = calc1stOrderDerivative1DGaussian(y_values, sigma_y)';
    im = imfilter(im, filter_y,'symmetric');
end

function [ D_values ] = calc1stOrderDerivative1DGaussian(x_values, sigma)
    % Getting the 1D Gaussian values
    G_values = calc1DGaussian(x_values, sigma);
    
    % Obtaining the 1st order derivative of the 
    D_values = -2.*x_values./2./sigma^2.*G_values;
end

function [ G_values ] = calc1DGaussian(x_values, sigma)
    G_values = 1/sigma/sqrt(2*pi).*exp(-x_values.^2./2./sigma^2);
end

Thresholding

Function	Description
imbinarize	Binarize 2-D grayscale image or 3-D volume by thresholding
graythresh	Global image threshold using Otsu's method
multithresh	Multilevel image thresholds using Otsu’s method

Morphological Operations

morphological operations

Function	Description
strel	Morphological structuring element
imdilate	Dilate image
imerode	Erode image
imopen	Perform morphological opening. Morphological opening is useful for removing small objects and thin lines from an image while preserving the shape and size of larger objects in the image.
imclose	Perform morphological closing. Morphological closing is useful for filling small holes in an image while preserving the shape and size of large holes and objects in the image.
bwskel	Skeletonize objects in a binary image. The process of skeletonization erodes all objects to centerlines without changing the essential structure of the objects, such as the existence of holes and branches.
bwperim	Find perimeter of objects in a binary image. A pixel is part of the perimeter if it is nonzero and it is connected to at least one zero-valued pixel. Therefore, edges of interior holes are considered part of the object perimeter.
bwhitmiss	Perform binary hit-miss transform. The hit-miss transform preserves pixels in a binary image whose neighborhoods match the shape of one structuring element and do not match the shape of a second disjoint structuring element.The hit-miss transforms can be used to detect patterns in an image.
imtophat	Perform a morphological top-hat transform. The top-hat transform opens an image, then subtracts the opened image from the original image.The top-hat transform can be used to enhance contrast in a grayscale image with nonuniform illumination. The transform can also isolate small bright objects in an image.
imbothat	Perform a morphological bottom-hat transform. The bottom-hat transform closes an image, then subtracts the original image from the closed image. The bottom-hat transform isolates pixels that are darker than other pixels in their neighborhood. Therefore, the transform can be used to find intensity troughs in a grayscale image.
bwmorph	Morphological operations on binary images. To perform morphological operations on a 3-D volumetric image, use bwmorph3.

Other Segmentation Operations

Function	Description
imclearborder	Suppress light structures connected to image border
watershed	Watershed transform

ROI Labeling / Feature Extraction

Function	Description
regionprops3	Measure properties of 3-D volumetric image regions. See also regionprops
bwconncomp	Find and count connected components in binary image
labelmatrix	Create label matrix from bwconncomp structure
bwlabeln	Label connected components in binary image

Transformation

Function	Description
imtranslate	Translate image
imrotate3	Rotate 3-D volumetric grayscale image

Registration

Function	Description
imregister	Intensity-based image registration
imregconfig	Configurations for intensity-based registration
imregdemons	Estimate displacement field that aligns two 2-D or 3-D images
imwarp	Apply geometric transformation to image

Visualization

Types of MATLAB Plots

General

Function	Description
figure	Create figure window
subplot	Create axes in tiled positions
tiledlayout	Create tiled chart layout for displaying subplots
nexttile	Create axes in tiled chart layout
gca	Current axes or chart
gcf	Current figure handle
axis	Set axis limits and aspect ratios
xlim	Set or query x-axis limits. See also ylim and zlim
xlabel	Label x-axis. See also ylabel and zlabel
title	Add title
sgtitle	Add title to grid of plots

2D

Function	Description
imagesc	Display image with scaled colors
contour	Contour plot of matrix. Good for showing boundary of ROIs in mask
plot	2-D line plot
xline	Vertical line with constant x-value. See also yline
scatter	Scatter plot
imshowpair	Compare differences between images
stackedplot	Stacked plot of several variables with common x-axis

3D

Function	Description
volshow	Display volume
orthosliceViewer	Browse orthogonal slices in grayscale or RGB volume
sliceViewer	Browse image slices
scatter3	3-D scatter plot
plot3	3-D point or line plot
view	Camera line of sight

Codes for Batch Analysis

%% MAIN SCRIPT (or in command window)
% path to data
ops.filedir = '../originals/'; % folder

% path to saving directory
ops.savedir = '../outputs/'; % folder

if ~exist(ops.savedir, 'dir')
    mkdir(ops.savedir)
end

loop_through_folder(ops);

disp('Done:)')


%--------------------------------%
%                                %
%       (LOCAL) FUNCTIONS        %
%                                %
%--------------------------------%

function loop_through_folder(ops)
    %% loops through folder and subfolders
    filelist = dir(foldername);
    ops.savepath = ops.savedir;

    
    for i = 1:length(filelist) % the first two are '.' and '..', skip
        
        if strcmp(filelist(i).name,'.')
            % % save data in the original data folder
            % ops.savepath = filelist(i).folder;
            % if ~exist(ops.savepath, 'dir')
            %     mkdir(ops.savepath)
            % end

        elseif strcmp(filelist(i).name,'..') 
            continue
        
        elseif filelist(i).isdir % is a folder
            disp(filelist(i).name);
            
            % create new folder for saving data for each subfolder
            ops.savedir = fullfile(ops.savedir, filelist(i).name);

            if ~exist(ops.savedir, 'dir')
                mkdir(ops.savedir)  
            end

            ops.filedir = fullfile(filelist(i).folder,filelist(i).name);
            loop_through_folder(ops);

        elseif contains(filelist(i).name, '.czi') % change file format
            ops.filename = fullfile(filelist(i).folder, filelist(i).name);
            [~,filename,~] = fileparts(ops.filename);
            
            % create new folder for saving data for each image file
            ops.savedir = fullfile(ops.savepath, filename);
            if ~exist(ops.savedir, 'dir')
                mkdir(ops.savedir)  
            end

            try
                custom_function(ops);
            catch
                warning('Error. Check file.')
            end
        end

    end
end

function your_processing(ops)
    %--------------------------------%
    %                                %
    %     CODES FOR PROCESSING       %
    %                                %
    %--------------------------------%
    tic;
    [~,filename,~] = fileparts(ops.filename);
    disp(filename);

    % Put your tested pipeline here.
    % You may remove the plots used for varifying the steps.
    % It is recommended to save the key plots for debugging purpose.
    % Add the line 'close(gcf)' to close the figure after saving.

    toc;
end

Downloading Add-Ons

Downloading official Matlab add-ons to Clarke require admin right. Let Jane knows if you need any additional toolbox. For other scripts that could be useful to your project, you may download them into /PATH/TO/PROJECT/readmes and add that to path using

addpath('/PATH/TO/PROJECT/readmes')

Setting up Git Source Control

For any github resources, it is a good idea to stay updated and keep track on the changes you've made to the scripts. Follow the step in here to set up set up source control.

Common lab SOPs:

Bioinformatics-related:

scRNA-seq processing

Image analyses-related:

Programming-related:

Useful BASH commands

8. Image and Signal Analysis

Useful Tools/Papers

Experiment Design

Calcium Image Analysis

Cell Segmentation

Spots Detection

Motion Tracking

ImageJ / Fiji Macro

Save as tiff

Export ROI as CSV

Matlab

Useful MATLAB functions

Starting a Script

Loading and Saving Data / Figures

The Bio-Formats Library

Loading

Saving

Creating File and Folder Names (string manipulation)

Data Types

Reshaping / Interpolation

Spatial Filtering and Intensity Adjustment

Temporal Filtering

Zero-padding

Baseline Detection

Edge Detection

Thresholding

Morphological Operations

Other Segmentation Operations

ROI Labeling / Feature Extraction

Transformation

Registration

Visualization

General

2D

3D

Codes for Batch Analysis

Downloading Add-Ons

Setting up Git Source Control

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Clone this wiki locally