replicateFest: An R Package and Shiny App for Analysis of T Cell Receptor Repertoire Data from the Functional Expansion

Imagine your immune system as a massive, bustling army. Inside this army are millions of soldiers called T cells, each wearing a unique badge (a T cell receptor, or TCR) that lets them recognize specific enemies. Sometimes, scientists want to find out which specific soldiers are ready to fight a particular enemy, like a virus or a cancer cell.

To do this, they use a special training exercise called a FEST assay. Think of it like a "mock battle" where they expose the T cells to a specific antigen (the enemy) and see which soldiers get excited and multiply (expand) in response.

The Problem: The "Copy-Paste" Confusion
The challenge is that these mock battles are often run multiple times (replicates) to make sure the results are real and not just a fluke. However, every time you run the experiment, there are tiny differences—like how the weather changes slightly between two days, or how a few extra soldiers might show up by chance.

Before this paper, scientists didn't have a good way to sort out which T cell expansions were real reactions to the enemy and which were just random noise caused by these small variations between experiments. It was like trying to hear a whisper in a room where the wind keeps changing direction; existing tools couldn't help them tune out the wind.

The Solution: The "replicateFest" Toolkit
The authors created a new digital toolkit called replicateFest. You can think of this as a smart, automated referee for these mock battles. It comes in two forms:

1. A computer program (R package): For scientists who like to write code.
2. A user-friendly website (Shiny App): For scientists who prefer clicking buttons and seeing charts.

How It Works
This referee uses two different strategies depending on the data it gets:

• If there's only one experiment: It uses a standard math check (Fisher's exact test) to see if the T cells expanded significantly compared to a control group (a "no-antigen" practice session).
• If there are multiple experiments (replicates): It uses a more advanced statistical method (negative binomial modeling). This is like a referee who watches the same game played five times and only calls a "goal" if the team scores consistently across all five games, ignoring the random bounces of the ball.

The toolkit then gives scientists a clear list of "winners"—the specific T cells that truly expanded because of the antigen. It also distinguishes between two types of winners:

• FEST-expanded: Soldiers that grew more in the antigen battle than in the control battle.
• FEST-positive: The elite soldiers that grew more than they did in any other condition tested.

Did It Work?
The team tested their new referee in two ways:

1. Fake Data: They created a "synthetic" dataset where they knew exactly which T cells were supposed to be the winners. replicateFest correctly identified them, proving the math works.
2. Real Data: They applied it to past data about HIV-1. The toolkit successfully found the same results as the original researchers, proving it can handle real-world data and help check the quality of past studies.

Where to Find It
The toolkit is free for anyone to use. You can download the code from GitHub or try the interactive website directly through the link provided in the paper. It's designed to make the complex job of analyzing T cell responses more reliable and easier to understand.