Image Analyst MKIIProtocols
HomeWorkflowImage Processing BasicsFunctions GlossaryPipelines GlossaryProtocolsQuick How ToSearch

Analyze Time Lapse Recordings with Image Analyst MKII

Video Tutorials

Publications using Image Analyst MKII

Try Image Analyst MKII for free

Download

Calculation of spectral unmixing coefficients for mitochondrial membrane potential measurement

Spectral crossbleed can be calculated from a double-stained sample, given that the two spatially co-localized fluorophores behave differently in time. However, more precise determination of the crossbleed is possible with using single-fluorophore stained samples and the same methods given below. Here the spectral crossbleed is calculated from the data set used for the potentiometric calibration.

For TMRM/PMPI time lapses without performing separate single fluorophore recordings:

  1. Open only the baseline and the compete depolarization section of a potentiometric experiment. Note: the aim is to open a time course where fluorescence intensities vary in time in a distinct manner between the two channels.

For single-fluorophore recordings:

  1. Perform points 2 to 8 for each single flurophore image (a single time point is sufficient) and construct the coefficient matrix as given in point 10.  For each fluorophore run take only the coefficients from the result of the “Spectral Unmix NMF Status” that is in the column corresponding to the channel number of the fluorophore, e.g. the value in the top right corner comes from the PMPI-only sample and the value in the bottom-left corner from the TMRM-only sample, where TMRM is channel #1 and PMPI is channel #2. Note: for increased precision depolarize plasma membrane potential before measuring PMPI crossbleed.

  1. Select the “Prepare two channel time lapse for intensity analysis” pipeline in the Pipelines/”Intensity Measurements” main menu point.

  2. The default parameters of the pipeline have been configured to work with this particular recording, otherwise set it up similarly to the pipeline used for image processing for the potentiometric calibration above.

  3. Press the  button on the main toolbar or in the bottom of the Multi-Dimensional Open dialog to load and process the images.

  4. Select the “Blind Spectral Unmixing with NMF” function from the Math menu.

  5. Optionally set the “Write result to coefficient matrix” to Yes, and type a new matrix name (e.g. TMRM/PMPI).

  6. Select both channels as Operand A (on the right of the function parameters) and press  in the toolbar. Note: the order matters here, TMRM has to come first, then the PMPI. In the operands list Image windows appear in the order of their creation.

  7. Review the images that the spectral unmixing is correct, and no gross oversubtraction happen by the end of the experiment (appearing as negative values or black shadows around or within objects).

  8. In pipelines requiring spectral unmixing, parameter named “Spectral Unmix Coefficient Matrix”, give either the above entered new matrix name, or copy the coefficients from the “Spectral Unmix NMF Status” window. Note: it is advised to run the calculation in multiple positions (the “Spectral Unmix NMF Status” also appears in the Excel Data Window) and calculate mean or median of coefficients.

  9. Enter coefficients into the “Spectral Unmix Coefficient Matrix” parameter of the pipeline. The matrix can be entered:

  1. As ={{1,”PMPI crossbleed into TMRM channel”},{“TMRM crossbleed into PMPI channel”,1}} by writing the measured coefficients instead the quoted expressions.

  2. As a matrix name, where the named matrix is entered in the Main menu/Edit/Preferences/”Convolution kernels, Matrices” tab.

Protocol by Akos A. Gerencser 07/29/2015 V1.0