transcriptomics / NSPC

Analysis of neural stem and progenitor single cell omics datasets
GNU General Public License v3.0
2 stars 0 forks source link

negative value in count matrix #1

Open priya11234 opened 1 year ago

priya11234 commented 1 year ago

Hi, I am very new to scrna analysis. I have downloaded the .h5ad object from synapse.org. and fetch the count matrix(code is mentioned below)

adata = read_h5ad('hNSPC.h5ad') sp.io.mmwrite("matrix.mtx",adata.X) or In R Convert("hNSPC.h5ad", dest = "h5seurat", overwrite = TRUE,verbose = T) data <- LoadH5Seurat("hNSPC.h5Seurat") counts_data <- Matrix::Matrix(as.matrix(data@assays$RNA@counts), sparse = T)

why count matrix have negative value ?

Thank you

daniel-d-liu commented 1 year ago

Hi Priya, negative values in the count matrix can arise after preprocessing steps. Usually, this would be from regressing out variables (not done for our dataset) or from batch correction of different sequencing runs (which we did using combat). If you'd like to start from a raw count matrix, we also uploaded an unprocessed .h5ad file to the Synapse. This matrix should only have values >=0. Keep in mind you will need to batch correct based on sequencing run to reproduce our results.