Analysis Pipeline - dattalab/moseq2-app GitHub Wiki
Overview
There are three ways to run the MoSeq analysis pipeline, which are Jupyter notebooks and CLI. The Jupyter notebooks and CLI can be accessed through installation using Conda or the MoSeq2 Docker Image. The fosllowing sections contain instructions for Jupyter notebooks and CLI. You can find more information about the directory structure and intermediate yaml files in the analysis pipeline here.
Jupyter notebooks Instructions
MoSeq2 Extract Modeling Notebook
MoSeq2 Analysis Visualization Notebook
Flip Classifier Training Notebook
Command Line Interface Instructions
Which one do I use?
Below is a comparison of the two main MoSeq2 interfaces: the command-line interface, and the Jupyter notebook. You don't have to choose just one, as they can be used interchangeably.
Jupyter Notebook
| Pros | Cons |
|---|---|
| Easy to use | Doesn't support automation |
| Shows both the code blocks and the output | |
| Comes with interactive widgets to analyze model results |
Command Line Interface (CLI)
| Pros | Cons |
|---|---|
| Can be used in bash scripts flexibly for automation and parallelization | Can be confusing for users that have never used a CLI |
| Limited visualization and interactive capabilities |
If you are interested in using the CLI for extraction and modeling, but using the interactive widgets in the Jupyter notebooks to find parameters and analyze results interactively, you can find more information in extraction and modeling CLI documentation and the extraction and modeling or analysis notebook documentation.
Useful Resources
Try Our Test Data
If you want to explore MoSeq functionalities, check out our test data.