feiyoung / ProFAST

Probabilistic factor analysis for spatially-aware dimension reduction across multi-section spatial transcriptomics data with millions of spatial locations
2 stars 1 forks source link

error when running iscmeb_run() or RunHarmonyLouvain() #4

Open SiaGuo opened 2 months ago

SiaGuo commented 2 months ago

I've ran the function, FAST(), to integrate 4 samples. But when I tried to run iscmeb_run() or RunHarmonyLouvain(), I got the error like, Fit SC-MEB2... Finish variable initialization K = 10, iter = 2, loglik= -289562.512426, dloglik=0.999865 Error: matrix multiplication: incompatible matrix dimensions: 2213x15 and 0x0, or the error like, **Use Harmony to remove batch in the embeddings from FAST... Error: element-wise multiplication: incompatible matrix dimensions: 100x4 and 100x1.

I wonder what's wrong with this, cause I've ran on another datasets and it worked well. So anyone has some ideas? Thank you!

feiyoung commented 2 months ago

You can enhance the quality control to prevent outliers from occurring when utilizing the CreatePRECASTObject function by adjusting the following arguments to higher values: premin.spots, premin.features, postmin.spots, and postmin.features.

SiaGuo commented 2 months ago

Thanks for your reply, I will try and let you know the results.


发件人: @. @.> 代表 Wei Liu @.> 发送时间: 2024年7月13日 07:56 收件人: feiyoung/ProFAST @.> 抄送: SiaGuo @.>; Author @.> 主题: Re: [feiyoung/ProFAST] error when running iscmeb_run() or RunHarmonyLouvain() (Issue #4)

You can enhance the quality control to prevent outliers from occurring when utilizing the CreatePRECASTObject function by adjusting the following arguments to higher values: premin.spots, premin.features, postmin.spots, and postmin.features.

― Reply to this email directly, view it on GitHubhttps://github.com/feiyoung/ProFAST/issues/4#issuecomment-2226543706, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AWIYTUIYWAH4YJ2GVXZZNTTZMBUKTAVCNFSM6AAAAABKWZF226VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRWGU2DGNZQGY. You are receiving this because you authored the thread.Message ID: @.***>