Guidance for Planning Single-cell/nucleus Experiments

Use this document to help guide you in thinking about the details of your prospective single-cell/nucleus experimental design. You can replace the text in the boxes below and save the file to PDF.

If you need help in answering questions in this document or would like to chat with one of our experts please reach out to the Research Informatics (RI) Bioinformatics group at ribhelp@msi.umn.edu.

Questions to Consider When Designing a Single Cell Experiment

This section lists important questions that should be addressed during the experimental design process.

1. What are the major biological questions you would like to answer with your single-cell/nucleus experiment?

2. Sequencing Library Type:

Are you doing a standard single cell or single nucleus RNA-seq library for measuring gene expression (aka GEX) or are you planning to use different library approaches, examples include: ATAC, Multiome, VDJ, ADT, HTO/CITE-Seq, etc.

3. Sequencing Library Vendor/Protocol:

Examples include 10X Genomics, Parse Biosciences, Illumina Single Cell, Drop-Seq, etc. There are also specific types of library protocols such as 10X Genomics Flex kit vs 3’ libraries.

4. What type of cells/tissue and organism?

Many experimental workflows only work for human or mouse samples or require fresh cells.

5. Will you capture single cells or single nuclei?

Single nuclei libraries may be more appropriate for certain cell types that are multi nucleated or have irregular shapes.

6. How many samples in total for this experiment: A sample is your unit of biological replication. If you are pooling cells or tissues from multiple donor organisms then a pooled sample is considered to be a single replicate.

7. Experiment Design  

A. Variables: Describe your experimental variables (e.g. treatment, condition, sex, age, time, etc.)

B. Replication: How many different groups (different possible combinations of variables) are present given those variables? How many biological replicates per group?

C. Batch effects: Will the samples be collected at different times? Will they be captured in separate captures or sequenced at different times?

8. Sample Preparation:

A. If the starting material is tissue, how will the tissues be dissociated (mechanical, enzymatic, combo)? Note: This can bias the frequencies of the cell types recovered.

B. Do these cells/tissues have any special characteristics that would influence their cell recovery in a microfluidics-based capture such as 10X? Some examples include: abnormally large size cells (e.g. adipocytes, neurons, skeletal muscle cells), stickiness of cell surface, extreme sensitivity to dissociation, and fibrotic tissues. These can all result in low cell or nucleus recovery.

C. How will these cells/tissues be stored prior to library preparation?  Will these cells/nuclei be fixed? [note: fixation is compatible with 10X flex and Parse Biosciences single-cell preps.]

D. Will an enrichment method (e.g. FACS, differential/density gradient centrifugation, magnetic bead enrichment) be applied prior to cell capture? If cells will be sorted by FACS, what markers will be used? Is the method/ kit used for enrichment compatible with the sequencing library kit (talk to staff at the UMGC if they will be preparing and/or sequencing the libraries). 

E. What cell types do you expect to observe within this sample after performing any above enrichment steps? (Think about cell types that might be contaminating as well.) 

F. If applying a treatment to some of your single-cell samples, do you expect a large difference in cell viability? (e.g. cells treated with chemotherapy will have lower viability.) 

9. Analysis Limitations: Rare Cell Types

A. What are the frequencies of cell types within this sample (e.g. 1% B-cells, 70% neutrophils)? 

B. What are the cell types you are most interested in within this sample?

C. Use the SCOPIT Shiny app (https://alexdavisscs.shinyapps.io/scs_power_multinomial/) to estimate the number of cells to sequence in a sample in order to recover low frequency cell types of interest.

Design Guidance and Considerations

This section highlights specific calculations, tools, and analysis considerations that are valuable for single cell/nuclei experiments.

1. Estimating the Ability to Capture Rare Cell Types

This section can be used to estimate how many rare cells you may be able to capture in your experiment.

A. Estimated # cells loaded for sequencing capture (this depends on capture type/vendor):

B. Estimated # multiplets given the estimated cells being loaded:

Note: You can obtain multiplet rate information from the documentation for the sequencing kits that you are interested in directly from the manufacturers:

C. Estimated # cells captured (the number of cells you are intending to capture):

D. Final Number of Cells  = Estimated # cells captured (#1A) minus estimated cell multiplets (#1B):

E. Estimated # of cells that might be captured from the rarest cell type of interest.

This is equal to the frequency of rarest cell x Final Number of Cells (#1D)

Note: Ideally would like > 100 cells for DEG comparisons involving this cell type.

F. If applying a treatment to some of your single-cell samples…

Do you anticipate large changes in cell type frequency due to the treatment (Y / N)? After treatment, do you think your cell population(s) of interest will still have > 100 cells?

2. Power Analysis for Multiple Samples

If this is a multi-sample experiment, please check out the scPower R package to perform power calculation estimations for differential gene expression analysis using various cell type data from published single-cell studies. There is an scPower web application Note: If the web app for scPower is down, the scPower point and click GUI can be accessed by installing the R package in RStudio and following the instructions on the Github: https://github.com/heiniglab/scPower

3. Data Analysis Considerations

A. If this is a multi-sample design without multiplexing (i.e. one sample per 10X capture), how do you plan to analyze the samples together downstream to mitigate batch effects from separate captures? 

B. Given the library type (GEX, ATAC), what strategies will you use to identify cell types in your experiment?

C. Do your cells have reporter genes (e.g. GFP) that you would like to be able to detect during your analysis and that are not in the reference genome sequence?

D. Do you plan to perform trajectory inference analysis (pseudotime or RNA velocity) on your single-cell/nucleus data?

4. General Considerations:

Final thoughts

After answering the questions in this guide, do you think you will be able to answer all the major biological questions you have with this data set? Do you have additional questions that you would like to discuss?

Please reach us out to us if you need additional support.

For questions about single-cell/nucleus capture, library prep, and sequencing (details and cost), contact staff at the UMGC (single-cell@umgc.umn.edu).

For questions about experimental design, single-cell data analysis, or contract bioinformatics analysis service (timeline/cost), contact ribhelp@msi.umn.edu.

You can save your work in this form to a PDF file, by clicking the export button below.

Return to top