Entering edit mode
12 months ago
sarahmanderni
▴
120
Hi,
Based on the literature, some genes like Gm42418, AY036118 and Malat1 are indicators of rRNA contamination or low quality cells and are suggested to be removed from the count matrix before normalization. I am dealing with a dataset that after removal of controversial genes (MT genes, Ribo, Gm42418, AY036118 and Malat1), a 30 to 40 percent of cells the remain to have very low library size (total UMI per cell), e.g. less than 500 total UMIs per cell. Will you filter out these cells as apparently the main content has been rRAN and MT genes? Are samples with high proportion of such cells reliable for the downstream analysis?
What species and types of cells was this data collected from? Some cell types and some species have higher levels of mito mRNA in scRNA-seq and thus elevated levels may not be indicative of low qual/apoptotic cells.
Refer to the following for more background/context:
Osorio, Daniel, and James J Cai. 2021. “Systematic Determination of the Mitochondrial Proportion in Human and Mice Tissues for Single-Cell RNA-Sequencing Data Quality Control.” Edited by Anthony Mathelier. Bioinformatics 37 (7): 963–67. https://doi.org/10.1093/bioinformatics/btaa751
“Removal of Dead Cells from Single Cell Suspensions Improves Performance for 10x Genomics Single Cell Applications.” 2017. CG000130 Rev A Technical Note. https://cdn.10xgenomics.com/image/upload/v1660261286/support-documents/CG000130_10x_Technical_Note_DeadCell_Removal_RevA.pdf
Thanks for the links! These are colon samples from human.
Rather than manually removing the cells expressing expression the indicators of rRNA contamination or low quality cells, I would suggest you to use standard approaches like SoupX or CellBender to deal with these type of data. You can check the expression of these genes in pre and post quality control data. Remember that you need to have both raw and filtered count matrices to analyze the data using these approaches.
Thanks for the suggestion. SoupX has been used to correct for the ambient RNA. However, the above mentioned markers are still among the top 5 highly expressed genes.