To put it shortly, CellexalVR is a virtual reality platform for the visualisation and analysis of single-cell gene expression data. Below follows some about the platform.
Any. SmartSeq2, 10X Chromium, it doesn't matter. You perform the initial analysis using your preferred pipelines/methods, and these are sent to for visualisation and further anlaysis.
Single-cell expression data is commonly reduced to 3-dimensions using some form of multi-dimensional scaling (MDS) such as principal components, tSNE, and pseudotime analysis. The resulting projections can have very complicated structures which are difficult to appreciate and understand using conventional plotting methods such as rgl plots in R. These plots have limited rotatability and zoom, so getting an overall picture of the MDS landscape is hindered. Selecting subgroups of the graphs for further analysis can be difficult especially without any programming knowledge. CellexalVR overcomes this by providing a full view and total maneuverability of the point cloud in an expansive VR environment. Selecting subgroups in the VR environment is no harder than grabbing them with your hands.
Another benefit of CellexalVR is that all data objects communicate with each other. For example if you have multiple representations of the same data you can cells between them.
A decently specced gaming PC with preferably an Nvidia GTX1080 graphics card. This is our current setup:
Yes, a multi-user feature has been added since version 0.9.0. Users can connect to the same session and share and collaborate with the same data being in the same physical room. Everyone who wants to connect to the same session must have the same dataset downloaded on their machines. See the manual for more information on how to set up a multi-user session.
Yes. CellexalVR 0.10.0 introduced the spectator view, which let's you join a multi-user session without the need of VR equipment. Interaction in this mode is limited.
CellexalVR is coded in Unity, and most backend calculations are performed in R by the cellexalvrR package which also needs to be installed.
The user guide will tell you how to format data for input, but we have made cellexalvrR, an R package which makes this easy. After you have performed your MDS analyses in R, create a cellexalvr object as instructed in the documentation and then call the output2cellexalvr function that will created the necessary files and database.
CellexalVR will not preprocess data. Mapping/normalisation/filtering and projection to 3d coordinates is all done by the user using their preferred methods. CellexalVR will pick up from this point onwards.
Calculations are performed in R during a CellexalVR session. Users are welcome to change or modify the R package to suit their needs and install their own version.
CellexalVR was created in the Computational Genomics Lab (PI: Dr Shamit Soneji) at the Division of Molecular Hematology and Stem Cell Center in collaboration with the VR labs at the Ingvar Kamprad Design Center, both at Lund University.