TissueMiner is a framework to perform a multiscale analysis of an entire developing epithelium acquired at cellular resolution over several hours or days. It includes (a) tools to build a database from a timelapse movie, (b) ready-to-use tools to analyze various aspects of cell and tissue morphogenesis, and (c) a powerful but yet easy to use R programming interface to allow for the implementation of custom analyses.
New TM release on October 2nd 2017:
-
New release is 1.0.3 see Version history for changes
-
TissueMiner must be run through DockerToolBox (Win7, Win10, MacOS) or Docker engine (Linux)
-
To update TM run
docker pull etournay/tissue_minerin a terminal (Linux) or in a Docker-QuickStart-terminal (MacOS, Windows) -
Please make sure that the tm alias in your .bashrc includes the
-u rstudiooption:
alias tm='docker run --rm -ti -v $(dirname $PWD):/movies -w /movies/$(basename $PWD) -u rstudio etournay/tissue_miner'
- CAUTION for Linux Users: this only works if your user ID is 1000.
- The Rstudio IDE is bundled with TissueMiner (many thanks to the Rstudio and rocker teams) to easily run the TissueMiner API and to create new scripts for addressing project-specific questions.
First, TissueMiner tracks cells in timelapse image series of a tissue and stores the results in a database. To do so, it uses TissueAnalyzer to segment and track cells in time and space.
Next, TissueMiner extracts and organizes information about cell geometry, topology (cell neighbor relationships) and ancestry from the TissueAnalyzer outputs into a database. By using sqlite, each movie is converted into a single database file.
TissueMiner comes along with a great variety of tools to work with the created movie database. These tools include
- Cell geometry analysis
- Cell topology analysis
- Cell lineage browsing
- Cell contributions to tissue deformation
Many of these tools provide tabular results as well as rendered movies to make the complex nature of timelapse more accessible.
TissueMiner provides a convenient R programming interface to query, quantify and visualize cell dynamics during epithelium morphogenesis in R. The functions provided with the TissueMiner API allow one to query the movie database for extracting quantities like:
- Cell state properties (position, area, anisotropy, cell packing geometry, fluorescence intensity)
- Rates and orientation of cellular events (divisions, cell neighbor changes, extrusions, shape changes)
- Rates of tissue deformation (area expansion and convergence-extension) contributed by each type of cellular event (tensorial description)
This API can then be employed to
- Visualize quantified data in graphs or directly on the movie images
- Do statistics (distribution of cell area, anisotropy, packing, bond length, vertex multiplicity, ...)
- Synchronize different movies in time
- Compare values between multiple movies and ROI's
- Do multiscale quantification and visualization using both tracked ROI's (Lagrangian description) and fixed grids (Eulerian description)
TissueMiner ships as a single Docker image and must be run using the Docker containerization platform.
-
Step1: Install Docker
- for Linux, please install the docker engine (and optionally you may have to fix DNS issues here).
- for MacOS and Windows systems, please install the docker toolbox (180Mb).
-
Step2: Download the TissueMiner docker image called etournay/tissue_miner (~700Mb)
- on Linux just open a terminal, on Mac or Windows open a Docker Quick Start Terminal
and type in the command below
- on Linux just open a terminal, on Mac or Windows open a Docker Quick Start Terminal
docker pull etournay/tissue_miner
- For troubleshooting in case of out-of-date operating systems click here.
-
To get started with TissueMiner we suggest to do Quickstart docker-TM-R Tutorials on our example data:
-
To get started with your own data analysis: docker-TM-R Tutorials
-
We provide the streamlined R scripts that are used in the tutorials to perform a simple analysis of a single movie. This includes a script entitled
analyse_movie.Rthat performs most of the analyses described in the Quickstart Tutorials. -
Build your own analysis in R (or Python) using the TissueMiner API:
- learn the necessary basics of the R language for TissueMiner
- learn the powerful libraries of TissueMiner for epithelium analysis using the exhaustive TM R User Manual with lots of examples, background information and API details
- You already know Python and you like to use it instead ? Go to the Tutorial in Python
-
For a more general overview consider our Resource Paper.
-
Commonly asked questions are answered in the FAQ section.
We provide four datasets:
- 1 demo sample (~35 Mb) to run Quickstart TM-R Tutorials
# In your unix terminal (or Docker QuickStart Terminal), type in:
curl http://bds.mpi-cbg.de/tissue_miner/demo.tar.gz | tar -zxvf -
- 3 big datasets (~800Mb each) to explore more advanced capabilities of TissueMiner described in the TM R User Manual
# In your unix terminal (or Docker QuickStart Terminal), type in:
# Dataset WT_1 (~800Mb)
curl http://bds.mpi-cbg.de/tissue_miner/WT_1.tar.gz | tar -zxvf -
# Dataset WT_2 (~800Mb)
curl http://bds.mpi-cbg.de/tissue_miner/WT_2.tar.gz | tar -zxvf -
# Dataset WT_3 (~600Mb)
curl http://bds.mpi-cbg.de/tissue_miner/WT_3.tar.gz | tar -zxvf -
Please use the github ticket system to report issues or suggestions. We also welcome pull requests.
If you like to use TissuMiner for your own research, please cite
Etournay et al. (2016). TissueMiner: a multiscale analysis toolkit to quantify how cellular processes create tissue dynamics. eLife, 5, e14334. doi:10.7554/eLife.14334
or
Etournay et al. (2015). Interplay of cell dynamics and epithelial tension during morphogenesis of the Drosophila pupal wing. eLife, 4, e07090. doi:10.7554/eLife.07090
RStudio is a registered trademark of RStudio, Inc. The use of the trademarked term RStudio and the distribution of the RStudio binaries through the images hosted on hub.docker.com has been granted by explicit permission of RStudio. Please review RStudio's trademark use policy and address inquiries about further distribution or other questions to [email protected].