This program was designed to interactively and flexibly slice datasets from functional imaging experiments along the time axis. A classical imaging experiment generates data sets of high dimensionality: The individual imaging trials, e.g. the response of a certain area to a stimulus is usually imaged as (x,y) images over time (t), resulting in 3d image stacks. Additionally, different stimuli (S) are given with a certain number of repetitions (R), resulting in 5 dimensional data (x,y,t,S,R).
This program serves the following purposes:
- It is a flexible data visualizer and inspector by offering interactive intensity scaling, color maps, overlays and data subselection to really see how your dataset looks like.
- It allows to select regions-of-interest (ROI) at which the data set is sliced along the temporal axis. The resulting time traces are directly displayed and updated upon any change.
- The traces can be extracted in
.csvformat, sorted to stimulus class S and repetition R for subsequent data analysis.
The program is still in a very early developmental stage, but already usable. Load your data from the load menu, trial labels can be added with the load labels function. Label files are expected to consist of a label in each line, with the index of each line corresponding to the respective data set. For example, if you want to load the files
- my_trial_1.tif
- my_trial_2.tif
- my_trial_3.tif
- my_trial_4.tif
and in your experiment trial 1 and 3 were stimulus A, 2 and 4 were stimulus B, then the label file has to look like
stimulus A
stimulus B
stimulus A
stimulus B
Fully supported are currently any .tif files, Zeiss .lsm files generated by 'pre-Zen' LSM 510 software (later ones might work) and .pst files from TILL Vision imaging rigs. Imports for other formats can be developed, please post a request and send me an example file.
To subset the currently displayed data set use the Data Selector on the top right. It supports intuitive ctrl+click, shift+click, ctrl+a etc. to select or deselect individual trials to display. The mode of display can be set with the centrally position icons, for example switch between dF/F raw display, average over frames, or a monochrome mode for individual trials with the raw data as the background optionally overlaid with the dF/F signal in a glow colormap.
Currently the datasets can be gaussian filtered along image and time axes, by convolving the x,y,t image stack with a gaussian kernel of size set in the options window. Frames that go into the background image for dF/F calculation can also be specified in the options window.
ROIs are added by simply clicking on the image, the type of ROI (circular or polygonal) can be selected in the extra Options window, opened by clicking on the icon in the toolbar. ROIs can be interactively resized and dragged to different regions, all traces will be updated interactively.
The ROI Manager supports the same multiple selection of ROIs just as the Data Selector. If multiple ROIs are selected, the coloring and traces display mode changes: The traces of all selected ROIs are now color coded to the ROI instead of the trial. This is useful if instead of comparing the response in two different trials, one wants to compare the response in two different areas which are within the field of view.
Depending on the selected datasets and the activated ROIs, the displayed data is sliced along the time domain under the area covered by the ROIs, each frames values are averaged. The sliced traces are color coded and can be displayed in two ways:
- All traces with a common time base
- sorted to stimulus identity
Both display modes can be accessed by selecting the corresponding tab of the traces visualizer. Both can also be detached from the main window by double-clicking on the tab, for example to move it on a separate screen. Traces are interactively updated upon any change to the ROIs or data selection.
After setting up all ROIs, the dataset can be sliced according to the pixels covered by the ROIs. The average value of those pixels for each frame is calculated and the resulting vectors can be written in different ways into .csv files:
.csv - normalwrites one file per loaded dataset, with the shape (t,ROI)..csv - sortedwrites one file per ROI and stimulus combination, with the individual columns representing the repetitions.
If a dt value is specified in the Options, the corresponding time vector is calculated and added to the first column of the .csv, with the stimulus start being set to 0, pre stimulus times are thus negative.
- Install the python 3.7 version of Anaconda for your operating system.
- Download the repostory
ILTIS - Open a terminal (Linux and Mac) or Powershell (Windows) and create an environment containing Python 3.7:
conda create -n ILTIS python=3.7(More on managing conda environments here) - Activate the conda environment created
conda activate ILTIS - Navigate into the repository containing ILTIS.
cd <path of the repository ILTIS> - Make sure that the file
setup.pyin your current path by looking at the output of the commandls(Linux/Mac) ordir(Windows). - Install ILTIS into the environment
pip install ./
- Open a terminal (Linux and Mac) or Powershell (Windows) and activate the conda environment created
conda activate ILTIS - Start ILTIS using the command "iltis"
iltis
- automatic calculation of ROIs (NNMF)
- movement corrections: rigid and nonlinear transformation based movement correction
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent