create FAQ section on wiki

[sreichl]

Informed by all previous talks, create an FAQ section on the wiki to address frequently asked questions. Collect here notes on that.

[sreichl]

Frequently Asked Questions (FAQ)

This FAQ is based on the interactive discussion following the talks and presentations of MrBiomics.

General & Technical Overview

What exactly is MrBiomics and what languages does it use?
MrBiomics is a framework of composable Snakemake workflows designed to automate bioinformatics for multi-omics analysis. While the framework is language-agnostic, the underlying scripts currently utilize R and Python.

How does the user interface work?
MrBiomics does not use a graphical user interface (GUI). Instead, users interact with the system by editing configuration files (config files) and using a command-line interface to call Snakemake, which executes the defined tasks from end-to-end.

Does MrBiomics require a high level of programming knowledge?
The system is designed for computational biologists of all levels. The primary requirement is the ability to use Snakemake from the command line, rather than developing new workflows. By removing the "engineering" hurdles—such as manual package installation and dependency management—it allows researchers to focus on the science rather than coding infrastructure.

Installation & Compatibility

How are software dependencies and "dependency hell" managed?
MrBiomics handles dependencies internally via Conda. Each task in a workflow triggers Snakemake to automatically look up, install, and activate the necessary Conda environment. This process is transparent to the user and ensures that the correct versions of all tools are used without manual intervention.

What are the hardware and server requirements?
Workflows can be run locally on Linux or macOS, or on high-performance computing (HPC) clusters/servers. While pre-processing large raw data is best suited for HPC parallelization, downstream analysis—such as re-running a count matrix after a parameter change—can often be completed locally in minutes or hours.

Is MrBiomics portable across different systems?
Yes. To ensure 100% portability and reproducibility, a top priority for the project is full containerization. MrBiomics supports Docker and other containers, which prevents workflows from breaking due to system-level updates.

Data & Workflow Management

How should I format my raw data for use in MrBiomics?
Specific formatting requirements for raw data are detailed in the project's documentation. Additionally, the repository provides example datasets so users can quickly compare their file structures to the expected input.

Can I access intermediate outputs if I run an entire "recipe"?
Yes. Every module within a recipe is independent and self-contained. Even when running a full end-to-end analysis, the system generates and stores full results for each individual module.

How does MrBiomics handle versioning and reproducibility?
Workflows are clearly versioned on GitHub. Users can stream specific versions of modules directly from GitHub using a URL, ensuring that an analysis performed today can be replicated exactly in the future.

Comparison & Integration

How does MrBiomics differ from other tools like Galaxy, scverse, or nf-core?

• scverse: While scverse focuses on package compatibility on a data-object basis, MrBiomics operates a layer above by assembling these packages into comprehensive, multi-step workflows.
• Galaxy: MrBiomics avoids the complexity and manual "button-clicking" often associated with Galaxy by using code-based, automated design decisions that have been validated in peer-reviewed publications.
• nf-core (Nextflow): Unlike some nf-core workflows that may repeat code across different analyses, MrBiomics emphasizes high abstraction and modularity, allowing the same code to be reused across different modalities (e.g., RNA-seq and ATAC-seq).

Can I integrate my own custom scripts or AI-generated code?
Yes. MrBiomics is modular, meaning you can plug in your own scripts or modules alongside established recipes. While AI can be used to generate code, MrBiomics advocates for using "best practice" workflows that have been validated by humans with real data to ensure scientific reproducibility.

[sreichl]

Frequently Asked Questions (FAQ)

💻 Usage & Requirements

Do I need to be an expert in R or Python to use MrBiomics?
No. While the internal scripts are written in R and Python, the framework itself is language-agnostic. MrBiomics is built on Snakemake, meaning you only interact with high-level configuration files (YAML/JSON) and the command line. The framework handles the underlying code execution for you.

Can I run this on my local computer, or do I need a cluster?
You can do both. MrBiomics runs on local Linux and macOS machines as well as HPC clusters (e.g., Slurm). For intensive preprocessing (raw reads to counts), a cluster is recommended for parallelism. However, downstream analyses (differential expression, enrichment) can often be run locally on a standard workstation.

How difficult is it for a wet-lab biologist to use?
The framework is designed to remove "engineering headaches" rather than biological thinking. While there is a learning curve to using the terminal and Snakemake, we have seen researchers with no prior cluster experience successfully run full end-to-end analyses (from raw data to pathway enrichment) within a single week.

---

🛠 Technical Implementation

How does MrBiomics handle "dependency hell" and package versions?
MrBiomics uses Conda to manage software environments internally. You do not need to install individual bioinformatic packages; Snakemake automatically creates, activates, and deactivates the required environments for each task. For absolute portability, we are also working on full Docker/Container support to ensure the environment remains stable across different operating systems.

Is the analysis reproducible if I share my project?
Yes. Every module is versioned on GitHub. By specifying a version tag in your configuration, you ensure that the exact same code is executed every time, even years later. This makes your research 100% technically reproducible.

How long does a typical analysis take to run?

• Preprocessing: Moving from raw sequencing reads to a count matrix typically takes about one day.
• Downstream Analysis: Once you have your counts, rerunning steps (e.g., to fix a typo in your sample sheet or tweak a plot) takes only minutes to hours.

---

🧬 Workflow Philosophy

How does this differ from other platforms like nf-core or Galaxy?
While many platforms focus on "one-size-fits-all" pipelines, MrBiomics emphasizes composability. We treat downstream analysis as modular. For example, you can swap an ATAC-seq quantification module for an RNA-seq module, but keep the exact same downstream code for unsupervised analysis and enrichment. This "Functional Knowledge Management" ensures that an improvement to one module benefits all your projects.

Do I have to run the "Full Package" every time?
No. You can use our pre-built Recipes for end-to-end analysis, or you can pick and choose individual Modules. If you already have a count matrix, you can skip the preprocessing and start directly with the differential expression or multi-omics integration modules.

Can I integrate my own custom scripts into the workflow?
Absolutely. MrBiomics is designed to be the "skeleton" of your project. You can easily plug in custom Python/R scripts or other specialized tools (like TOBIAS for footprinting) alongside the standard modules to answer your specific biological questions.