A single-cell sequencing project has different parameters: the number of samples, cells per sample and reads per cell. But how do you decide how many cells you need? Or how deep you need to sequence? Here, we explain what your choice should be based upon.
Number of samples
The first single-cell sequencing parameter that you should decide on is your number of samples. How many samples you need is highly dependent on your biological question. If you expect a lot of cell types, are looking for a rare cell type, or compare groups, you might need to include a bit more samples to answer your question.
We also advise you to hand in duplicates as a backup.
Cells per sample
With “cells per sample,” we mean the number of cells per sample that ends up in your single-cell data. This is a number that you aim for. So, the number of cells per sample you collect before the experiment is a lot higher since cells will be lost during processing steps in the lab and cut-offs in data analysis.
Most projects aim for a number of cells between a few hundred to 10,000 cells per sample. But how do you decide how many cells you need?
The first consideration is the technical conditions. What platform are you going to use? How many cells can you collect initially?
We offer 10x Genomics projects with an aim from 3,000 – 10,000 cells per sample, which requires at least half a million cells in duplicates to start processing.
SORT-seq is available from 384 cells (one plate) and up, where you FACS sort the cells in your own lab into the plate. So first, determine the number of cells that are technically possible and fit the platform you chose.
The second consideration is your biological question. How many cells do you need to answer it?
Do you want to identify many cell types in your sample, or do you expect a rare cell type to be present? Then it would be best if you aimed a bit higher. Are you interested in a specific cell type that you will enrich for, with FACS for example, you don’t need that many cells.
The sequencing depth, the number of raw sequencing reads per cell, is closely related to the number of cells per sample. The number of reads usually varies between 30,000 and 150,000 per cell.
Generally, the higher your sequencing depth, the more genes you can identify. A high sequencing depth might be necessary when looking for a rare cell type or a low expressed gene, for example.
But when your cell types are easy to identify with not that many genes, you don’t need to sequence that deep.
Finally, we always recommend looking at the chosen parameters in existing publications. Since single-cell sequencing is becoming increasingly popular, the chances are high that some have done single-cell sequencing on a similar tissue.
You can find our publications on our website or use the many publications listed by 10x Genomics.
Want to discuss your single-cell sequencing parameters with us? Please book a meeting with one of our specialists.