How do i split data (scRNA)

Support Questions
1

Hello, thank you for your appreciation. i have two questions and I’m glad to assist.

  1. I have three normal sample files and three patient sample files. These are 10X platform data consisting of barcodes, features, and matrix files.
    I am trying to analyze them by integrating them. Here, I want to check a specific gene by dividing it into two groups, normal samples and patient samples. How can I split this?

  2. I am a different person from the one in the first question. Unlike the samples in question 1, my data is composed of scRNA data from multiple samples. The sample groups are: tumor core, tumor border, and adjacent non-tumor region. It is available on (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi.)
    However, instead of having barcodes, features, and matrix files configured for each sample, all samples are integrated into one set of barcodes, features, and matrix.

How can I divide the group into three groups, tumor core, tumor border, and adjacent non-tumor region, in this case?

Thank you.
Sincerly,
Mr Kim

2

i resolved the first problem. that was in the user guide.

But, still have a problem in 2.

Thank you

3

Hi Mr Kim,

From this comment, it sounds like the files need demultiplexed. instead of having barcodes, features, and matrix files configured for each sample, all samples are integrated into one set of barcodes, features, and matrix

@Ger_Bioinformatics Can you please point Mr Kim towards the instructions & code for demultiplexing these files?

Thanks!

4

Hi, this blog post should do. The main issue with demultiplexing is to figure out how the sample information is encoded. One of the most common approaches is to encode it in the column names of the count matrix, which is what this post is written around.

You may need to adapt the first part. Instead of loading the data with the readRDS function, you may need to use Seurat::Read_10x("path/to/your/sample/folder") (assuming the sample folder contains features.tsv/barcodes.tsv/matrix.mtx files.)