Dear all, I have a question about single cell RNA seq and ATAC seq. We want to conduct a single cell RNA seq for comparing the type of cell between group A, group B and group C ( A, B, C were the three types of the disease.). Also, we preforem the ATAC seq between the three groups.

How we could calculate the sample size for ATAC seq and single cell RNA seq, respectively?

Notably, we have found a paper titled "SCOPIT: sample size calculations for single-cell sequencing experiments" . But it provide a procedure for prospectively planning single-cell experiments or retrospectively evaluating if sufficient numbers of cells have been sequenced. (Not the number of patients）

How we could calculate the sample size for the number of patients for ATAC seq and single cell RNA seq, respectively? ( or We don't need to calculate it, we only need to calculated the number of cells)

Thank you for your help.

In my experience, any sample size calculation tool will tell you that you need very many samples, and usually that is again limited by both available money and feasibility of the experiment itself. As for any experiment, if you can, I would go for biological replicates, three at minimum, more if possible. scRNA-seq (at least 10x) can do cell hasing so you could combine several biological reps into the same library prep to save some money. ATAC-seq not laborious, as said, n=3 or more depending on available material and money. Does that make sense?

Dear ATpoint , Thank you for your suggestion which gives us clues about the sloution. It's very important for us. I am major in epidemiology and statistics, so I alaways help colleagues to revised the protocal of study design and statistics method including calculating the sample size. How to estimated the sample size for sevral high-throughput sequencing-based experiments or other experiments (such as, DNA, cfDNA, GWAS, single cell RNA, ATAC seq, RNA-seq, Epigenetic Sequencing，ChIP-seq, metabonomics，Microbiome)? I have learned some reaserch method paper. But the methods are not consistent. I always to ask myself honestly, Is it really necessary to calculate the sample size for these experiments?