Questions for running split-seq docker

[dmsalsgh97]

Hi, first. thanks for developing a wonderful pipeline for scRNA-seq!

I'm trying to use docker image on docker hub it works pretty well, but it always exited, even I used -itd option. Is there something wrong with my trials?

I'm trying to use the docker image with r-studio and jupyter-notebook for integrating split-seq data for Seurat and scanpy..

[yjzhang]

It might be because I haven't updated the image in 2 years... Anyway I just pushed a new image. Maybe it will work now?

[dmsalsgh97]

Umm, I tried it, but -it option seems not working on your docker image. When I run your docker image, the container exits with the following output..

Anyway, your code is working well for SRR6750042 with chemistry v1. Thanks for reply!

[yjzhang]

Okay, so it turns out that you need a --entrypoint argument because the Dockerfile has a predefined entrypoint:

docker run -it --entrypoint /bin/bash ayuezhang27/split-seq-pipeline

[dmsalsgh97]

Oh it works!

thanks for your kindness

[worker000000]

thanks a lot. I also want to ask a question about the docker version.
1 the option -chemistry seems to comes from the wet experiment, some how should I set that, the splitseq also has this? I see it in 10x genomics 2 in the part of Running the pipeline, it has 2 fatsq, why has 4 sample names? how the names comes, and wells like A1:B6 and other, how can I get such information

3 you have a part of Merging Sublibraries into a Single Matrix, so does it mean I ran each paired fastq through the part of Running the pipeline, and then doo step Merging Sublibraries into a Single Matrix,?

4 can this processed data be easily passed to seurat? if can, has you tried which version of seurat( it has 4 versions now). and what file should be passed through.

[dmsalsgh97]

Hi worker,

1. chemistry option determines the barcode sequences, etc. of split-seq protocol So, you must know which chemistry version the data was produced.

2. In my think, sample_name is for further analysis(like spatial, or multiplexing) using BC1 and Well position relationships.

4. you can use the output data for Seurat. when you run split-seq-pipeline all, there are 3 output files

by using R Matrix package, you can set the format proper for Seueat. below is my R script.

library(Matrix) library(Seurat)

dat = readMM('your file location for DGE.mtx') colnames(dat) = read.table('your file location for genes.csv',sep=',',header=TRUE)$gene_name rownames(dat) = read.table('your file location for cell_metadata.csv',sep=',',header=TRUE)$cell_barcode dat <- t(dat) #For transpose DGE matrix for seurat

Seurat_test = CreateSeuratObject(dat, names.delim='-', min.cells = 3, min.features = 200)

[worker000000]

thanks a lot. can you help me with questions3? 2 I see the supplymentary methods of the article of split-seq, it has many prepare steps of fastq data, is there any tool for doing this?

whcih version of seurat are you using? 3 or 4

[dmsalsgh97]

Hi worker,

The authors uploaded computational methods on this Github page, and by running the pipeline with all mode, you can get the gene counts for each cell (N x k matrix) for further analysis like Seurat.

I used the 3.1.5 version of Seurat.

[worker000000]

thanks

[dmsalsgh97]

Well, it seems like your docker disk mount seems making the problem '/sc/fastq_test//singlecells... ' And this is the Github page because Alexander B. Rosenberg is one of the contributions...

[worker000000]

thanks a lot. 1 after reading some paper, I do not find it takls about --chemistry v2 , do you know how to set this option. I think we can ignore it. 2 sample seems to just be important in split-seq, do you know the split-seq paper well position of the samples