Pytorch requirement

[maohub2008]

The pytorch >1.13.1 <2.0 is no longer available for downloading. I couldn't install the v0.1.0. Do you still have such restriction for your version 2.0?

[maohub2008]

Can you remove such restriction? I couldn't install the older version of the Pytorch.

[WPZgithub]

Hi @maohub2008 , sorry for this issue. I will update the code soon.

[maohub2008]

Thanks. I manually removed the restrictions from your source code. It worked well. I am using Pytorch 2.3.1 and CUDA 12.0. There are no compatibility issues.

Just some suggestions. Your method is a great tool. Since I am new to the bioinformatics world, more explanation in your tutorial will be very helpful for new birds like me. That will help the adoption of your CEFCON greatly.

For example, on the R code portion:
temp_R_dir = 'temp_nestorowa16_results'
os.makedirs(temp_R_dir, exist_ok=True)
cf.utils.prepare_data_for_R(adata, temp_R_dir, reducedDim='X_umap', cluster_label='leiden')
cf.utils.process_Slingshot_MAST_R(temp_R_dir, split_num=4, start_cluster=start_cluster, end_cluster=end_cluster)

If you can explain temp_R_dir and how to set it up on the user's computer, that will be very helpful.

You also didn't set the values for start_cluster and end_cluster in your sample code.

Thanks.

[maohub2008]

In your sample code, you mentioned that "Here, we directly use the trajectory information obtained from Beeline, and then call the R script in python to obtain differential expression information of genes along each trajectory." But in your code, I didn't see the Beeline was used. Can you explain it more? That will be very helpful, potentially in your tutorial too. Thanks.

[WPZgithub]

We used Beeline's benchmarking datasets, including the provided trajectory information, and we didn't call beeline's interface.

Thank you very much for your suggestions, I submitted some updates including the addition of new help information, I will be refining more help information! For the current help on how to use it please refer to the jupyter notebook files.