Raw Data examples

[MorselHoarder]

Hello,

I was wondering if you would be able to provide raw data examples for the train and prediction folders. It would be helpful to see the exact files and file structure that UNITO_Train_Predict/main.py is expecting.

You also mention in the paper that you provide an example script to retrieve the binary label output from the OMIQ platform. However, I don't see it anywhere in the repository. Would you be able to provide this as well?

Lastly, it would be nice if there were training- and prediction-ready csv files available for the user to load into the Google Colab notebooks to train their own model. The way it is now, it leaves the user to guess what data to put in and so most people give up. There's also no example model for users to try out on example data in the prediction-only Colab. I attempted to train a model using the mass cytometry data uploaded to Pennsieve, but I keep getting errors related to the data not being in the correct format. If we had an example data file like 01.T1_normalized.csv that was set up to match the gating_strategy.csv, then I think that would help clarify.

Thank you,
Andrew

[KyleeCJ]

Hi Andrew,

Thank you so much for your suggestions! That is very helpful! I realized that the data we uploaded is the direct output retrieved from OMIQ where the labels are stored in categorical format, basically the data format used in the framework is trying to use the hierarchical gating structure to convert the categorical label to binary label (by combining all the child gates to a single parent gate). I have prepared a shared link that stores the example data that can be used for training and prediction, and I will also update that in the repository. Thanks!

https://drive.google.com/drive/folders/13nlftDt0VgpCTHZoRiw8SSPFyoOnxOXL?usp=sharing