Tools for processing and analyzing data from behavior and imaging experiments. The package contains processing methods for data that is specific to the experiments I'm doing, but can be rewritten to work with any experimental system, there's a few places where user-specific functions need to be re-written (which are discussed in the documentation).
The experimental backbone is Rigbox, which uses an experiment management program called "Timeline". Behavioral processing is designed to work with experiments operated by the vrControl package in conjunction with timeline. These are virtual reality experiments where subjects (mice) run on a virtual linear track to collect rewards.
The imaging processing is designed to work with the standard output of suite2p, so if you are looking for a system to help analyze 2P data then you'll only need to change a few things.
Finally, the package assumes Alyx
database structure. This means that the data should be organized in a
directory first by mouse, then by date, then by session ID. These files should
be contained in a directory, which you can set in the
fileManagement
module with the localDataPath method.
There are two ways to install the vrAnalysis package. The first is by working
directly from the code after cloning the repository from GitHub. First,
navigate to whatever folder you want to store the cloned repository in. Then,
clone the repository, make a new conda (or mamba!) environment, and install
the requirements via the environment.yml file.
git clone https://github.com/landoskape/vrAnalysis
cd vrAnalysis
conda env create -f environment.yml # try mamba instead of conda if you haven't yet
Alternatively, you can do a pip install from GitHub if you don't want to edit the code at all. For this method, note that you have to choose which features you want to include. The core requirements exclude some packages required for reprocessing deconvolved calcium traces and exclude the red cell GUI.
In an existing conda environment, type:
pip install git+https://github.com/landoskape/vrAnalysis # core requirements
Or, for the extra installs, use one of the following. The registration component includes some packages that require special compilers. The gui component includes some plotting and interactivity related packages that I've found can conflict with other packages if you're using jupyter.
# pip install git+https://github.com/landoskape/vrAnalysis[registration] # include the registration packages
# pip install git+https://github.com/landoskape/vrAnalysis[gui] # include the red cell GUI
# pip install git+https://github.com/landoskape/vrAnalysis[all] # include everything
vrAnalysis can do the following things:
- Manage a database (can work with any SQL database with a few adjustments).
- Register and preprocess experimental data including behavior and imaging (will include facecam data soon).
- Manually annotate results, especially related to whether cells express a counter-fluorophore (finished) and for annotating tracked ROIs (not coded yet).
- Analyzing data using organized analysis objects that make it easy to do an analysis across groups of sessions from the database.
To add a new record to a database, check out this GUI! Note that the labels
and datatypes aren't hardcoded, so if your SQL database has a different
structure, this GUI will reflect it. For more info, go
here.
To figure out what sessions you have in your database that haven't been
registered (e.g. preprocessed), you can do this:
This isn't even close to a professional package so the documentation is primarily oriented towards helping future me remember what I've written. To that end, the documentation files contain a commentary on how to use the repository with lots of code blocks indicating standard usage. So, hopefully that's useful to anyone trying to use the repository or at least just to see what using this repository will look like.
Documentation Pages