The iSEE package provides an interactive user interface for exploring data stored in SummarizedExperiment objects. This repository contains the code required to construct the tours described in the iSEE paper by Rue-Albrecht et al. (2018).

Note: these tours can be resumed from within their respective iSEE application instance, using the question mark icon in the top-right corner of the Shiny application ( ).

Each tour is composed of three files:

• data.R: a script that contains the instructions to pre-process the data into a SummarizedExperiment or SingleCellExperiment object ready for iSEE.
• tour.txt: a set of step-wise instructions attached to various UI elements in the iSEE user interface.
• app.R: a script that uses the object generated by data.R to configure the iSEE application and launch the tour.

To launch a tour, successively execute data.R and app.R.

In addition, each tour is accompanied in this repository by two additional files illustrating the distribution of preconfigured iSEE applications in the form of Docker images, published on the Docker Hub:

• Dockerfile: command-line instructions for docker build to assemble the image.
• docker.R: wrapper for app.R (see above) that is executed by the Docker container on startup; it launches the preconfigured app on port 1234 of the container.

The port 1234 of a container can be published to any available port of the host machine using the -p option of docker run. For instance, docker run -p 1234:5678 makes the Shiny app available at http://localhost:5678 for the host machine.

Please refer to the Docker Documentation for further details.

The data.R script uses the Allen data subset in the scRNAseq Bioconductor package. It performs a small number of pre-processing steps (e.g., normalization, dimensionality reduction) before saving a serialized SingleCellExperiment object to file.

Live demo: https://marionilab.cruk.cam.ac.uk/iSEE_allen

Docker demo:

docker pull kevinrue/isee_allen
docker run -i --rm -p 1234:1234 kevinrue/isee_allen

The data.R script fetches the TCGA data set from the Bioconductor ExperimentHub in the form of an ExpressionSet. It converts this to a SingleCellExperiment object and performs a small number of preprocessing steps (e.g., PCA, t-SNE) before saving the object to file.

Note: the first time that the script is run, it may take a few extra minutes, as it downloads and caches a copy of the data set if you haven't one already. Subsequent runs of the script will launch the tour significantly faster, as they will use the locally cached data set. Refer to the documentation of the ExperimentHub for further details.

Live demo: https://marionilab.cruk.cam.ac.uk/iSEE_tcga

Docker demo:

docker pull kevinrue/isee_tcga
docker run -i --rm -p 1234:1234 kevinrue/isee_tcga

This is somewhat more involved as the relevant data, while publicly available, need to be processed and analyzed. This can be achieved by following these steps:

1. Download and unpack the PBMC 4K dataset from the 10X Genomics website.
2. Run the analysis script. Modify the fname variable according to the path to the unpacked PBMC data files.
3. Move the generated sce.rds object into tours/ and run pbmc4k_app.R.

Live demo: https://marionilab.cruk.cam.ac.uk/iSEE_pbmc4k

Docker demo:

docker pull kevinrue/isee_pbmc4k
docker run -i --rm -p 1234:1234 kevinrue/isee_pbmc4k

The data.R script downloads a preprocesed version of the data set from Bodenmiller et al (2012). See the HDCytoData package for more information about how the data was processed.

Live demo: https://marionilab.cruk.cam.ac.uk/iSEE_cytof

Docker demo:

docker pull kevinrue/isee_cytof
docker run -i --rm -p 1234:1234 kevinrue/isee_cytof