PyCaret - LORIS Tutorial

[paulocilasjr]

This tutorial guides users in building and evaluating an ICB treatment prediction model using the LORIS dataset and Galaxy-PyCaret. It covers model comparison, performance assessment, and tests the robustness of the LORIS Score.

Changes:
/topics/statistics/images/loris_tutorial
/topics/statistics/tutorials/loris_model

The images were generated by me.

[bgruening]

@paulocilasjr I would like to get at least one more reviewer looking at it content wise. I have pinged people in our Matrix channel.

If you need this urgendtly, we can add draft: true like here and merge it.
This means you can reach the training via URL but it will not be listed on the frontpage.

[anuprulez]

I will find sometime this week and have a look :)

[anuprulez]

@paulocilasjr thank you for the tutorial. I have added a few comments inline. I tried to run the tutorial on https://cancer.usegalaxy.org/ but am unable to. I upload the datasets as TSV files and try to run the PyCaret Model Comparison tool but the jobs go to the paused state. I have tried two times, each time verifying the input datasets. Do I miss something? Please refer to the following screenshots:

[anuprulez]

The LORIS LLR6 model should be discussed in the tutorial. Details such as model architecture, features etc should be mentioned. It would improve the tutorial.

[paulocilasjr]

Thank you. I have included more information regarding the LLR6 model.

[paulocilasjr]

Regarding the jobs being paused, I had the system administrator check the job you started. He said that it was in the queue to run; there was nothing wrong with the job itself, just a long queue time.

Could you try again and let me know if you are still facing long queue times?

@paulocilasjr thank you for the tutorial. I have added a few comments inline. I tried to run the tutorial on https://cancer.usegalaxy.org/ but am unable to. I upload the datasets as TSV files and try to run the PyCaret Model Comparison tool but the jobs go to the paused state. I have tried two times, each time verifying the input datasets. Do I miss something? Please refer to the following screenshots:

[anuprulez]

The input datasets went into an error state and I think because of that jobs stay paused. I am wondering why the uploaded dataset file went into an error state

[anuprulez]

I tried again with the steps mentioned in the tutorial. I think there is some issue with importing datasets. I upload them as seen in Step1 and don't change the datatype of the dataset.

Step1

Step2

Step3

[paulocilasjr]

@anuprulez
Thank you for the screenshots; they helped me trace the problem.

Solution:
When uploading the data (Step 1), leave the type option as Auto-detect and everything should work just fine.

Let me know if you encounter any issues trying this.

Explanation of why you are getting the error:

• In Step 2 (see your figure), I noticed something unusual compared to my session. When selecting the datasets in the tool, the end of the selected filename shows this: 21: Chowell_train_Response.tsv (as csv).
• This issue originates from Step 1 (see your figure), where the Type option was set to TSV instead of Auto-detect.
• You might have discovered a bug. I’ve already reported it to the system administrator.
  Thank you for pointing this out.

[paulocilasjr]

I added a tip just in case.

[anuprulez]

I could run the tutorial by importing datasets in the "auto detect" mode for extension. It works fine and looks good. I have one minor suggestion:

Check that the data format assigned for the file is **tabular**.

Can you make this change for avoiding confusion related to the dataset format? The dataset's path has .tsv extension but Galaxy shows it as tabular after upload and not as tsv. Therefore, I think informing user to be aware when they see that the dataset does not have a tabular extension after upload would be enough. Thanks!

[paulocilasjr]

Thank you.

[anuprulez]

Looks good to me! thanks @paulocilasjr