Dynema (Dynamic eQTL mapping for single cells)

Dynema is a method to map single-cell eQTL effects at real cellular resolution.

Dynema's generalized framework enables testing complex regulatory effects including:

• Main effects: A standard eQTL effect, independent of any context
• Interaction effects: eQTL effects that change depending on one (single-context) or multiple (multi-context) contexts
• Total effects: Joint effect of main and interaction eQTL components. This effect captures any genetic signal driven by either main or interaction eQTL effects

Dynema scales to genome-wide analysis, accounts for repeated measurements (multiple cells per donor), and provides calibrated p-values by using cluster-robust variance estimators (CRVEs). Additionally, it provides robust inferences in extreme scenarios such as small number of donors via optional adaptive score bootstrapping (built on WildBootTests.jl).

System requirements

• Julia >= 1.9 (see installation)
• Software dependencies and their versions can be found in the Project TOML.
• Dynema V1.0.0 has been tested againt linux distributions (Red Hat Enterprise Linux v9.8 and CentOS v7) and MacOS Sequoia v15.7.7.

Installation

Dynema requires Julia >= 1.9 (easily installed with juliaup).

You don't need to write any Julia code to run Dynema!

The bin/ folder ships a self-contained command-line interface (dynema-map) that installs Dynema and all dependencies for you.

curl -L https://github.com/immunogenomics/Dynema.jl/archive/refs/heads/main.tar.gz | tar -xz
mv Dynema.jl-main Dynema
cd Dynema
./bin/dynema-map --help

Dynema's installation takes ~2–3 min on a personal laptop.

Demo and instructions for use

For more details and a quic-start demo tutorial (~ 1-2 minutes to run), see the documentation website.

Pre-print

You can check our pre-print on BioRxiv for more details and usage cases. The Julia API code used for this manuscript can be found here https://github.com/immunogenomics/Dynema_analysis for reference.