edeno / spyglass

Neuroscience data pipeline for reproducible research built by Loren Frank Lab at UCSF

spyglass

The Frank lab Datajoint pipeline facilitates the storage, analysis, and sharing of neuroscience data to support reproducible research. It is designed to be interoperable with the NWB format and integrates open-source tools into a coherent framework.

Setup

Installing packages

1. Clone this repository:

   git clone https://github.com/LorenFrankLab/spyglass.git

2. Set up and activate a conda environment from environment.yml:

   cd spyglass
   conda env create -f environment.yml
   conda activate spyglass

3. Install this repository:

   pip install -e .

Setting up database access

1. Ask Loren or Eric to set up an account for you on the Frank lab datajoint database. Note that you have to be connected to UCSF LAN to access this server.

   If you're not affiliated with UCSF or if you are just looking to try out spyglass, then you will need to set up a different MySQL server. For example, you can set up your own local server with a Docker image of a MySQL server configured for Datajoint (see instructions and/or tutorial notebook).

2. Add the following environment variables (e.g. in ~/.bashrc). We assumes that you are interacting with the database on a computer that has mounted stelmo at /stelmo (if the mount location is different, change accordingly). For this to take effect, log out and log back in, or run source ~/.bashrc in the terminal.

   export SPYGLASS_BASE_DIR="/stelmo/nwb/"
   export SPYGLASS_RECORDING_DIR="/stelmo/nwb/recording"
   export SPYGLASS_SORTING_DIR="/stelmo/nwb/sorting"
   export SPYGLASS_WAVEFORMS_DIR="/stelmo/nwb/waveforms"
   export SPYGLASS_TEMP_DIR="/stelmo/nwb/tmp"
   export KACHERY_DAEMON_HOST="typhoon"
   export KACHERY_DAEMON_PORT="14747"
   export KACHERY_STORAGE_DIR="/stelmo/nwb/kachery-storage"
   export KACHERY_TEMP_DIR="/stelmo/nwb/tmp"
   export FIGURL_CHANNEL="franklab2"
   export DJ_SUPPORT_FILEPATH_MANAGEMENT="TRUE"

   Note that a local SPYGLASS_TEMP_DIR (e.g. one on your machine) will speed up spike sorting, but make sure it has enough free space (ideally at least 500GB)

3. Check if you have access to the kachery daemon. Open up a terminal, activate the conda environment, and type

   kachery-client info

   If it says you do not have access to the daemon, ask Kyu to add your username to the kachery-users group.

   If you're running your own kachery daemon, then this doesn't apply.

4. Configure DataJoint. To connect to the Datajoint database, we have to specify information about it such as the hostname and the port. You should also change your password from the temporary one you were given. Go to the config directory, and run dj_config.py in the terminal with your username:

   cd config # change to the config directory
   python dj_config.py <username> # run the configuration script

   Again, if you're using your own MySQL server, then you may need to change the other settings as well.

Finally, open up a python console (e.g. run ipython from terminal) and import spyglass to check that the installation has worked.

Tutorials

The tutorials for spyglass is currently in the form of Jupyter Notebooks and can be found in the notebooks directory. We strongly recommend opening them in the context of jupyterlab.

• 0_intro.ipynb: general introduction to the database
• 1_spikesorting.ipynb: how to run spike sorting
• 2_curation.ipynb: how to curate spike sorted data