A GUI tool for easy and smooth visualisation and analysis of Spatial Transcriptomics datasets
Clone or download
Pull request Compare This branch is 1 commit ahead of jfnavarro:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
assets
cmake
data
src
.gitignore
AUTHORS
CHANGELOG
CMakeLists.txt
DEPENDENCIES
Dockerfile
LICENSE
README.md
REPLACE_QT_LIBRARIES
_clang-format
ubuntu_dep_and_compile.sh

README.md

Spatial Transcriptomics Research Viewer

The ST viewer is a tool that allows the visualization of spatially resolved gene expression data on top of HE stained tissue figures with the correct location.

The ST Viewer is cross platform which means that it can be built and run in OSX, LINUX and WINDOWS.

The ST viewer allows to interact with the data in real time. Users can see where specific genes are expressed and how expressed they are. It has different thresholding, normalization and visualization options. It also allows to select areas of the tissue to obtain gene patterns to later do DEA or spot classification using machine learning.

The ST viewer uses the data generated with the ST Pipeline https://github.com/SpatialTranscriptomicsResearch/st_pipeline, which consist of a matrix of counts in TSV format where genes are columns and spot coordinates are rows in the fllowing form:

eg. 1x2

Where 1 represents the X coordinate and 2 represents the Y coordinate.

The ST viewer also requires a tissue HE image and an optional 3x3 alignment matrix (to convert array coordinates to image pixel coordinates).

Note that the referred 3x3 aligment matrix file must have the following format:

a11 a21 a31 a21 a22 a23 a31 a32 a33

If the HE image is cropped to the array boundaries then no alignment matrix is needed.

The ST viewer allows to pass a spot coordinates file to correct the coordinates positions and/or to only show the spots under the tissue. This file is compatible with the output format of the ST Spot Detector https://github.com/SpatialTranscriptomicsResearch/st_spot_detector

If you want to load a dataset you can go to the "Datasets view" and click in the button "Import dataset" then a dialog form will be shown where you can load the matrix of counts, the HE image and other files. You can also import a dataset automatically if all its files are inside a folder with the option "Load folder" or you can use a meta-file to load a dataset. The meta-file must describe where all the dataset's files are and it should have the following JSON format:

{
    	"name": "test",
    	"tissue": "test_tissue",
    	"species": "test_species",
    	"comments": "test_comments",
	"data": "/Users/user/test_dataset/stdata.tsv",
	"image": "/Users/user/test_dataset/image.jpg",
	"aligment": "/Users/user/test_dataset/alignment.txt",
	"coordinates": "/Users/user/test_dataset/spots.txt",
	"size_factors": ""
}

After that you can just double click in the dataset to open it. (more detailed information about this in the wiki).

You can use our public datasets hosted in http://www.spatialtranscriptomicsresearch.org/ if you want to try the ST Viewer.

Authors

Read AUTHORS file

Dependencies

Read DEPENDENCIES file

Manual

See Wiki

License

See LICENSE for the license terms and DEPENDENCIES for the 3rd party libraries that are used in this software.

Contact

For any question/bugs/feedback you can contact Jose Fernandez Navarro jose.fernandez.navarro@scilifelab.se

Binaries (Install from binaries)

Binaries(installers) for MAC and Windows are provided under the tab releases. Before you install the ST Viewer trough the binaries you must do the following (in case you have not done it already): The binary provided for MAC and Windows require that you have installed the same R version as the one used to generate the binary (indicated in the releases tab)

OSX
  • Download and install R from https://cran.r-project.org/ (in case you do not have it already)

  • Open R and install the following packages (Rcpp, RInside, RcppArmadillo, DESeq2, edgeR, Rtsne and SCRAN)

      source("https://bioconductor.org/biocLite.R")
      biocLite("DESeq2")
      biocLite("scran")
      biocLite("edgeR")
      install.packages(c("RcppArmadillo", "Rcpp", "RInside", "Rtsne"))
    
  • Download the installer (DMG) open it and drag the ST Viewer icon to Applications and then the ST Viewer will be installed in your system.

Windows
  • Download and install R from https://cran.r-project.org/ (Use the 32 bits option)

  • Download and install Rtools 32bits from https://cran.r-project.org/bin/windows/Rtools/

  • Open R and install the following packages (Rcpp, RInside, RcppArmadillo, DESeq2, edgeR, Rtsne and SCRAN)

      source("https://bioconductor.org/biocLite.R")
      biocLite("DESeq2")
      biocLite("scran")
      biocLite("edgeR")
      install.packages(c("RcppArmadillo", "Rcpp", "RInside", "Rtsne"))
    
  • Make sure that your PATH environment variable contains Rtools' bin, Rtools MinGW's bin and R's bin paths

      eg PATH=C:\RTools\3.4\bin\;C:\RTools\3.4\mingw_32\bin\;C:\Program Files\R\R-3.4.3\bin\i386
    
  • Make sure that you do not have another MinGW in your PATH variable

  • Make sure to have a environment variable called R_HOME pointing to where R is installed (its root folder)

      eg R_HOME=C:\Program Files\R\R-3.4.3
    
  • Download the Windows installer double click on it and follow the instructions, once done the ST Viewer will be installed in your system.

If you have problems running the ST Viewer on a windows machine, make sure that R is properly installed/updated, that it is accesible by all the users, that the required R packages are installed and functional, that the R_HOME and PATH variables are configured correctly and ultimately that the visual studio redistributable libraries are installed in your system (although, this should not really cause any problem).

Docker container

This has been tested on a Linux Ubuntu 17.04 as host system.

Build the image:

docker build . -t st_viewer

Launch the image, mounting also the volume where you have the dataset. For example if your dataset is in /home/user/STDatasets/ you need to launch with the option -v /home/user/STDatasets:/STDatasets which you can then find, via the fileBrowser of the Viewer in the directory /STDatasets. Note that you need to allow the root user to use your Display to see the Viewer:

xhost +local:root

Then launch the image according to where your file are located.

docker run -d -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix:rw -v /home/user/STDatasets:/STDatasets st_viewer

Building from the source

  • Download and install CMake 3.7.2 (Latest versions of CMake 3.9.x and 3.8.x have compatiblity issues with Qt so I recommend to download and install the version 3.7.2 or a previous one)

    Tips (for Linux and OSX; first download the file to a folder then in that folder open a terminal):

      wget https://cmake.org/files/v3.7/cmake-3.7.2.tar.gz
      tar -xvzf cmake-3.7.2.tar.gz
      cd cmake-3.7.2
      ./configure
      make -j4
      sudo make install
    

    For Windows you can download the installer from https://cmake.org/files/v3.7/cmake-3.7.2-win64-x64.msi (64 bits) or https://cmake.org/files/v3.7/cmake-3.7.2-win32-x86.msi (32 bits). Remember to add CMAKE to the system path when asked.

  • Download and install Qt open source from http://qt-project.org/downloads (Choose Desktop application and Open Source and then use the defaultsettings and location). For Windows you must choose the mingw32 option and include QT Charts.

  • Download and extract QCustomplot from http://www.qcustomplot.com/release/1.3.2/QCustomPlot.tar.gz

  • Download and build Armadillo from http://arma.sourceforge.net/download.html

    NOTE (Armadillo only needs to be built for Linux and OSX, for Windows you just need to download and extract it to a folder):

    • Download the latest stable release and then open a terminal and type (x.xxx.x refers to the Armadillo version):

        tar -xvf armadillo-x.xxx.x.tar.xz
        cd armadillo-x.xxx.x
        ./configure
        make
      
  • Download and install R from https://cran.r-project.org/ (in case you do not have it already) (For Windows use the 32 bits option)

  • Download and install Rtools 32bits (Only for Windows) from https://cran.r-project.org/bin/windows/Rtools/

  • Open R and install the following packages (Rcpp, RInside, RcppArmadillo, DESeq2, Rtsne and SCRAN)

      source("https://bioconductor.org/biocLite.R")
      biocLite("DESeq2")
      biocLite("scran")
      biocLite("edgeR") 
      install.packages(c("RcppArmadillo", "Rcpp", "RInside", "Rtsne"))
    
OSX
  • Make sure that XCode and XCode Command Line Tools are installed (check by typing "xcode-select" on a terminal) If needed to you can install them from the Apple store (https://developer.apple.com/xcode/). I recommend to update to the latest version of XCode.

  • Clone the repository to a specific folder and build the application

      git clone https://github.com/jfnavarro/st_viewer.git
      mkdir st_viewer_build
      cd st_viewer_build
      cmake -DCMAKE_BUILD_TYPE="Release" -DCMAKE_PREFIX_PATH="/path/to/libraries" -DCMAKE_OSX_SYSROOT="/path/to/macosx.sdk" -DCMAKE_OSX_DEPLOYMENT_TARGET=version ../st_viewer
    

    Where :

    DCMAKE_BUILD_TYPE = indicates the type of building ("Debug" or "Release" which is the default)

    DCMAKE_PREFIX_PATH = the path to where Qt, armadillo and qcustomplot are installed

    eg: "/Users/username/Qt/5.9.2/clang_64;/path/to/qcustomplot;/Users/username/armadillo"

    DCMAKE_OSX_SYSROOT = provides the path to the MacOS X SDK that is to be used (Only OSX users)

    eg: /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.13.sdk/

    DCMAKE_OSX_DEPLOYMENT_TARGET = use 10.12 or 10.11

  • Compile the application

    make -j4 
    
  • Run the application by clicking on the STViewer.app icon that can be found in

      /st_viewer_build
    
Linux

If you are on a Ubuntu, you can use the ubuntu_dep_and_compile.sh to install the dependencies and compile the code with:

sudo ubuntu_dep_and_compile.sh

Or you can follow the process outlined below.

  • Issue the following commands (Ubuntu, for Fedora you must use yum)

      sudo apt-get install git ubuntu-dev-tools
      sudo apt-get install libglu1-mesa-dev freeglut3-dev mesa-common-dev
    
  • Clone the repository to a specific folder and build the application

      git clone https://github.com/jfnavarro/st_viewer.git
      mkdir st_viewer_build
      cd st_viewer_build
      cmake -DCMAKE_BUILD_TYPE="Release" -DCMAKE_PREFIX_PATH="/path/to/libraries" ../st_viewer
    

    Where :

    DCMAKE_BUILD_TYPE = indicates the type of building ("Debug" or "Release" which is the default)

    DCMAKE_PREFIX_PATH = the path to where Qt, armadillo and qcustomplot are installed

    eg: "/Users/username/Qt/5.9.2/gcc;/Users/username/qcustomplot;/Users/username/armadillo"

  • Then type the following to build and install

      make -j4
      make install
    
  • To execute type :

      STViewer
      or
      /path/to/bin/STViewer
    

Note that for Linux you may want to update your LD_LIBRARY_PATH variable to contain the R and QT paths

eg: LD_LIBRARY_PATH=/usr/lib/R/lib/:/home/username/Qt/5.9.2/gcc_64/lib
Windows
  • Download and install Git for windows from https://git-scm.com/downloads

  • Open the GIT terminal and clone the repository :

      git clone https://github.com/jfnavarro/st_viewer.git
    
  • Make sure that your PATH environment variable contains Rtools' bin, Rtools MinGW's bin and R's bin paths

      eg PATH=C:\RBuildTools\3.4\bin\;C:\RBuildTools\3.4\mingw_32\bin\;C:\Program Files\R\R-3.4.3\bin\i386
    
  • Make sure that you do not have another MinGW in your PATH variable

  • Make sure to have a environment variable called R_HOME pointing to where R is installed (its root folder)

      eg R_HOME=C:\Program Files\R\R-3.4.3
    
  • Open a windows terminal (cmd.exe)

  • Type the following

      mkdir build
      cd build
      cmake -G "MinGW Makefiles" -DCMAKE_SH="CMAKE_SH-NOTFOUND" -DCMAKE_BUILD_TYPE="Release" -DCMAKE_PREFIX_PATH="/path/to/libraries" -DARMADILLO_PATH="C:\\armadillo" ../st_viewer
    

    Where:

    DCMAKE_BUILD_TYPE = indicates the type of building ("Debug" or "Release" which is the default)

    DCMAKE_PREFIX_PATH = the path to where Qt and qcustomplot are installed

    eg: "C:\Qt\5.9.1\mingw53_32;C:\qcustomplot"

    DARMADILLO_PATH = indicates where armadillo was extracted

    eg: "C:\armadillo"

    ../st_viewer = is the path where the ST Viewer was cloned/downloaded

  • Now build and install the ST Viewer by typing (you must run this as an administrator):

      mingw32-make install
    
  • By default the ST Viewer will be installed in "Program Files" but that can be changed with the CMake variable -DCMAKE_INSTALL_PREFIX (it is recommended to install the ST Viewer as an administrator)