nasqar / seuratv3wizard

A web-based interactive (wizard style) application to perform a guided single-cell RNA-seq data analysis and clustering based on Seurat v3

SeuratV3Wizard: R Shiny interface for Seurat single-cell analysis library

New Features:

Sep 5, 2019

• all plots can be downloaded
• file with multiple samples can be uploaded
• tsne and umap are both always calculated

Apr 25, 2019

• Seurat v3
• sctransform Option
• UCSC Cellbrowser

Online/Live instance:

You can try it online at http://nasqar.abudhabi.nyu.edu/SeuratV3Wizard

Pre-print:

NASQAR: A web-based platform for High-throughput sequencing data analysis and visualization

Run using docker (Recommended):

Make sure Docker (version >= 17.03.0-ce) is installed.

docker run -p 80:80 aymanm/seuratv3wizard

This will run on port 80

To run on a different port:

docker run -p 8083:80 aymanm/seuratv3wizard

This will run on port 8083

Local Install:

Make sure to have devtools installed first

devtools::install_github("nasqar/seuratv3wizard")

Optional: For ucsc cellbrowser support, make sure to follow the installation instructions here. For linux-based OS, type the following in the terminal:

sudo pip install cellbrowser=0.5.6

Run:

library(SeuratV3Wizard)
SeuratV3Wizard()

This will run on http://0.0.0.0:1234/ by default

---

To run on specific ip/port:

ip = '127.0.0.1'
portNumber = 5555
SeuratV3Wizard(ip,portNumber)

This will run on http://127.0.0.1:5555/

Screenshots:

Acknowledgements:

• Rahul Satija, Andrew Butler and Paul Hoffman (2017). Seurat: Tools for Single Cell Genomics. R package version 2.2.1. https://CRAN.R-project.org/package=Seurat

• Satija Lab

• Christoph Hafemeister, Rahul Satija (2019). Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression

• UCSC Single Cell Browser

seuratv3wizard's Issues

Keeps asking for barcodes.tsv.gz

The online version cannot detect that the barcodes.tsv.gz file is already in the upload and quits during file upload complaining that the file is missing.

"File upload job was stopped prematurely"
Error: barcodes.tsv.gz missing.

It only sees "features.tsv.gz" and "matrix.tsv.gz". It does not even recognize the file "bracodes.tsv.gz" has been added to upload list.

I hope this error does not continue into the docker version, which I am trying.

Improvement: select the cells according to nCount_RNA in the "VlnPlot (Filter Cells)" tab.

Hello,
it's me again.
Another suggestion for improvement:
To select the cells according to nCount_RNA in the "VlnPlot (Filter Cells)" tab.
Filtering on nFeature_RNA eliminates heterotypic doublets (2 different cell types, which express different genes: nFeature_RNA and nCount_RNA are higher).
Filtering on nCount_RNA eliminates heterotypic doublets, and also homotypic doublets (2 cells of the same type, which express the same genes: nCount_RNA is higher but nFeature_RNA remains unchanged).

Display potential biases on the umap/t-SNE

Hello,
Thank you for creating this tool. I would like to present it during a course for biologists.
However, I have a problem in the analysis and I want to offer you an idea for improvement.

I do not understand how to know if there is bias to correct. In general, we do a first run without regression of bias, then we display the potential biases on the umap/t-SNE to see if the cells were separated according to this potential bias (in this case it will be necessary to regress this bias) or not (not need to regress it).
For example, here we cannot see the influence of mitochondrial genes on dimension reduction.

In short, I would like to see the groups of genes created during the "QC & Filter", on the umap/t-SNE.
(Or maybe it can be added in the given object to the UCSC CellBrowser, to be display on it.)

In addition, another possible improvement would be the addition of the cell cycle assessment by seurat: https://satijalab.org/seurat/v3.1/cell_cycle_vignette.html

Thank you!

Bug (Improvement?): Cells by cluster in UMAP part?

Hello,
The section to get the list of cells in each cluster is placed in the t-SNE part. But it would also be necessary for UMAP (especially since this result does not depend on the non-linear dimensional reduction).
Thank you!

Can't browse cellbrowser, stuck at Prepared cellbrowser directory

Hello,

I successfully installed SeuratV3Wizard with its optional cellbrowser dependency, everything is working fine locally.

But then I tried to launch the app through a singularity image containing shiny-server and R 3.6.3 (From: rocker/shiny:3.6.3). The image contains also cellbrowser. And the app is launched as following :

singularity exec singularity_image.simg Rscript app. R

with app.R

appDir <- system.file('shiny', package = "SeuratV3Wizard"); shiny::shinyAppDir(appDir)
Everything works fine until I try to "Generate cell browser data", it gets stuck at Prepared cellbrowser directory /tmp//pbmcellbrowser and never pass to the following step "Converting cellbrowser directory to html/json files" as I get when i run it locally.

Do you have any idea to get through this issue ?

Thanks you so much

Clément

Combining Two or More 10X Runs

Hi nasqar,
Is there a way for me to merge/add more than one 10X run?
Plus, how to resolve "Maximum upload size exceeded"?

Bug draw PCA ICA

Hello,
In the "PCA Plot" tab, when we have done the ICA and we change the axes of the ICA, it also changes the axes of the PCA.

Viz Markers does not show list of genes to plot from

This issue is related to FindAllMarkers and showing top n genes.

Bug in file with multiple samples

Hello,
When the script uses "read.csv()", it converts "-" to "." in colnames, so when it makes the seurat object it ignores the sample names. So we cannot use multiple samples.
Maybe the separation between cell names and sample names should be "_" to solve the problem.

error in startup

Hi,
When i start up this app, it give me above error.

how can i fix this?
any advice would be thanks.
Best,
hanhuihong