- Beginners Guide (Training Resources)
- Data Analysis
- Expert Cytometry (ExCyte)
- 10 Things Smart Scientists Do Before Sorting Cells
- 3 ways to maximize your S/N ratio
- 4 Critical Components In Cellular Proliferation Measurement
- 5 Flow Cytometry Errors Reviewers Despise
- 5 Important Questions To Answer Before Submitting Your Data
- 5 Mistakes Scientists Make When Measuring Proliferation
- 6 Flow Cytometry Gating Tips That Most Scientists Forget
- 7 Advanced Flow Cytometry Data Analysis Tips
- 7 Mistakes People Doing Cell Cycle Analyses Make
- 8 Sample Prep Mistakes To Avoid
- 8 Time-Saving FACSDiva Software Tips
- Critical Steps in DNA Cell Cycle Analysis
- FMO Controls For Multicolor Flow Cytometry
- How Measure, Analyze, And Publish Apoptosis Data
- How To Differentiate T-Regulatory Cells By Flow Cytometry
- How To Get Funding For A Flow Cytometer Or Cell Sorter
- How To Perform A T-Test
- If You Don't Know This About FITC And PE...
- The Difference Between Efficiency, Recovery, And Yield
- Understanding The Jablonski Diagram
- When To Use (And Not Use) Flow Cytometry Isotype Controls
- Why Recovery Is The Best Measure Of Cell Sorting

- Flow Cytometry Training Courses
- For Beginners
- History of Flow Cytometry
- Online Training Resources
- Symposia
- Tips for Good Flow Cytometry Experiments
- Which Fluorochrome Combination?
- CYTO University
- Immune Cell Guide for Human and Mouse Antigens

- Core Facilities, Communities & Societies
- Flow Cytometry Instruments, Software & Upgrades
- Meetings & Conferences
- Online Resources
- Panel Design Resource
- Positions Available
- Reagents, Consumables & Services
- Science News
- Sponsors' Portals

# How To Perform A T-Test

ExCyte is a professional flow cytometry consulting company. Our mission is to provide expert instruction and to help leaders in the field connect and grow. We specialize in providing education opportunities to anyone utilizing flow cytometry or imaging in their research.

ExCyte’s courses provide a complex understanding of flow cytometry, instantly improving the attendees’ knowledge in experimental design, instrumentation, compensation, data analysis, and figure preparation. Trained experts in the field of cytometry provide unbiased lecture style training in basic and advanced topics in cytometry at a reasonable cost to the researcher.

### How to Perform a T Test

With the ability to capture expression data at the single cell level through many thousands of cells in a short time, flow cytometry data is very numbers rich. The importance of those numbers and how to use them in hypothesis testing is critical to ensure the robustness of the analysis.

After establishing the null hypothesis for the experiment, the type of statistical test, and the numbers necessary will become obvious. For example, if the null hypothesis states that the ‘treatment of B cells with thiotimoline does not change the expression of CD221B in normal patients.’

Based on this null hypothesis:

1. Assuming a gaussian distribution, a T-test will be performed.

2. The median and robust Standard Deviation will need to be known for the untreated and treated samples.

3. Additionally, the threshold needs to be establish. In this experiment the threshold is set to 0.01.

Here is the data:

Using a statistical tool, such as Graphpad Prism, the T-Test can be performed. Based on that result, the P value is less than 0.0001, which is below our threshold. Therefore the data is significant.

There are some limitations with directly comparing the median fluorescent intensities, especially in the case of a multi-center study with different instrumentation and such. In that case, the conversion of the data to a resolution metric, such as the RD is recommended. In this case, a one-sample T-Test can be performed on the data.

The RD is defined as follows:

RD = (MedianA-MedianB)/(rSDA+rSDB)

If the above data were to be converted, A would be defined as the treated sample and B as the control sample. That would result in the following data:

A one-sample T-test allows the comparison of a single sample to be compared to a known population mean. In this case, the mean would be 0. This is because the assumption is that the treatment causes no effect, so MedianA-MedianB would be 0.

Running the T-test, the results show, as expected, that this data is statistically significant.

The advantage of the RD is that issues with the experimental system (labeling, cell number etc), and the instrument can be smoothed out – making comparisons easier to interpret and understand.

This is a simple case of the power of statistics in flow cytometry. Please join me for an upcoming lecture where this and other topics such as statistical errors (and how to avoid them), ANOVA and linear regression will be discussed.

This entry was posted in Statistics And Data Analysis on February 21, 2014.