tdw1221 / scadie Goto Github PK

Hello,

Thanks for developing this great package. I followed the "SCADIE_vignette.Rmd"(https://github.com/tdw1221/SCADIE/blob/main/vignettes/SCADIE_vignette.Rmd)
and successfully run through the example data. Then I applied SCADIE_vignette.Rmd to my own data. It went through
ihc_bulk_groundtruth_scad_output <- Iterate_W_H_full_general( )

but generated error at the last step: ihc_bulk_groundtruth_sd <- Estimate_sd_general(input_list=ihc_bulk, update_W_method="SCAD" ,method="jackknife",cores=1,bs_num=NA,H_update_method="NNLS",H_update_gene="all",signature_gene_row_index=ihc_bulk$signature_gene_row_index,max_itr = 100,cutoff = 10^-6,jk_subsample = 15 )

The error info is: "Error in update_W(H_input = H_tmp, Y = bulk_expr_full) :
Duplicated or constant cols/rows in Y matrix, please check!
Calls: Estimate_sd_general ... mclapply -> lapply -> FUN -> SIR_itr_general -> update_W
Execution halted"

I checked the input data ihc_bulk, there is no duplicated or constant cols/rows in case, control, and all the data. Therefore, it seems that the error occurred in the middle of updating process.
I wonder if you know the possible reasons to cause the error and how to fix them.

Additionally, I wonder if it is OK to supply FPKM values as the input bulk expression data matrix.

Thanks a lot!

Hello,
My name is Saurabh Bhardwaj, and I am currently working on reproducing the results from your SCADIE publication, specifically focusing on the simulated data mentioned in the paper. I am seeking your guidance on the following query:

• In the publication, on page 4, step 5 discusses the calculation of z-scores and p-values for testing differential expression between pairs of entries W1ˆ ij and W2ˆ ij. I have reviewed the GitHub vignette provided, but unfortunately, I am unable to find the script corresponding to this step. It would be immensely helpful if you could share the script or guide me on where to find it.
  I greatly appreciate your time and assistance in addressing this query. Your support is crucial for the successful reproduction of the results, and I am eager to ensure accuracy in my work.

Thank you very much for your consideration, and I look forward to your response.

Thanks for developing this great package! I was able to replicate the method with my own data up through the second last step: Iterate_W_H_full_general().

When using Estimate_sd_general(), I get the following error. I have checked the input list and it should all be in the correct format and I've changed the number for "jk_subsample" to match my own data. Can you advise on what else might be causing this issue?

Error in seq.default(1, n_sample) : 'to' must be a finite number
In addition: Warning message:
In min(ncol(input_list$bulk_full_1), ncol(input_list$bulk_full_2)) :
no non-missing arguments to min; returning Inf