How to use WormID.org

1. Convert to NWB

Start by converting your data to the NWB format.

• For automatic ID, raw data should include a neuroPAL structural volume stack and/or a neural activity timeseries of volume stacks as well as associated metadata. Specific metadata requirements can be found in the tutorials

• Follow our data conversion tutorials and guidelines at github.com/focolab/NWBelegans

• NWB_tutorial.ipynb contains a step-by-step tutorial for data conversions

• Other examples of data conversions from .tif and .mat formats can be found in NWB_convert.py

2. Extract neural activity

Use eats-worm or another processing pipeline to segment neuron centers and extract neural activity traces from activity time-series volume stacks.

• Use eats-worm software available at github.com/focolab/eats-worm

• This pipeline is targeted at immobilized worms currently, not suitable for free moving recordings.

• See examples/NWB_example for an example of running an NWB dataset through the eats-worm pipeline

• You will likely need to tune some parameters based on the size and resolution of your data. Descriptions of parameters can be found in the ReadMe

• Review and curate automatic outputs from the algorithm and save to the NWB file with updated segmentation and activity traces

• If using another software, follow guidelines in NWB_tutorial.ipynb to add segmentation and activity extraction results to the NWB file

3. ID neurons

Use NeuroPAL_ID software or another tool to segment neuron centers and identify neuron cell-type labels from neuroPAL structural volumes

• Download pre-compiled NeuroPAL_ID app from github.com/Yemini-Lab/NeuroPAL_ID/releases/tag/v2.0b

• This app can be used to interface with NWB files with no coding experience

• Follow the instructions and tutorials at yeminilab.com/neuropal  for an explanation of how to use the software

• Run auto-segmentation of neurons, then remove false positive and add neurons that were missed by auto-segmentation

• Run the auto ID algorithm to get a preliminary set of assignments

• Follow the guidelines in the NeuroPAL manual found on the Yemini lab website to manually verify and curate these ID labels

• Every so often, you can rerun the auto ID algorithm to get updated assignments based on the ones you have already verified

4. Upload to DANDI

Upload data to DANDI to store on the cloud and share with the research community

• Start by making an account on the dandiarchive 

• After making an account, create a new dandiest and add a title and description

• Step-by-step instructions for uploading NWB files to DANDI can be found at dandiarchive.org/handbook/13_upload/

• You can edit metadata associated with your dataset directly on the Dandiarchive

• You can share this dandiset with others either by sharing the link or just the six-digit dandi ID

5. Add data to WormID corpus

Submit your dataset to the WormID corpus

• Once your data has been properly converted to the NWB format and uploaded to DANDI, you can submit it to the data corpus 

• Fill out this form to request adding your data to the corpus

• We will review your dataset and either request changes, or add it to the corpus of data available on this website

6. Experiment with analyses and visualizations

With your data converted to NWB, you can now run your data through code to test out various analyses and visualizations

• Code is at github.com/focolab/NWBelegans/tree/main/Analysis

• Refer to the bioRxiv preprint for examples of analyses and visualizations possible with this code, such as:

  • plotting point clouds for a single worm from an NWB file

  • visualizing a statistical atlas of neuron positions and colors

  • analyses and plots of cell positional variability, ganglion membership, synaptic connectivity, and cell lineage