-
Notifications
You must be signed in to change notification settings - Fork 5
Open
Description
Hi,
I can run fastAutoSmCCNet on the example data, but haven't managed to run it on any other matrices (real or random). Same error each time, which I haven't been able to trace back. Any insights would be much appreciated.
Reprex with random matrixes:
n_rows <- 7756
n_cols <- 10 #trialled a few values here, all errored
# Generate random data for both matrices
data1 <- matrix(runif(n_rows * n_cols, min = 0, max = 200), nrow = n_rows, ncol = n_cols)
data2 <- matrix(runif(n_rows * n_cols, min = 0, max = 20), nrow = n_rows, ncol = n_cols)
row_names <- paste0("GEN", seq_len(n_rows))
col_names <- paste0("samp_", seq_len(n_cols))
dimnames(data1) <- list(row_names, col_names)
dimnames(data2) <- list(row_names, col_names)
X <- list(data1, data2)
str(X)
List of 2
$ : num [1:7756, 1:10] 188.09 9.11 105.62 178.48 110.29 ...
..- attr(, "dimnames")=List of 2
.. ..$ : chr [1:7756] "GEN1" "GEN2" "GEN3" "GEN4" ...
.. ..$ : chr [1:10] "samp_1" "samp_2" "samp_3" "samp_4" ...
$ : num [1:7756, 1:10] 12.67 10.42 18.5 5.52 7.38 ...
..- attr(, "dimnames")=List of 2
.. ..$ : chr [1:7756] "GEN1" "GEN2" "GEN3" "GEN4" ...
.. ..$ : chr [1:10] "samp_1" "samp_2" "samp_3" "samp_4" ...
#mimic structure for Y
values <- sample(c(0, 1), size = n_cols, replace = TRUE)
Y <- matrix(values, nrow = n_cols, ncol = 1)
colnames(Y) <- "Pheno"
rownames(Y) <- paste0("Samp_", seq_len(n_cols))
result <- fastAutoSmCCNet(X = X, Y = Y,
preprocess = TRUE,
Kfold = 4,
subSampNum = 80,
DataType = c('rna', 'protein'),
CutHeight = 1 - 0.1^10,
saving_dir = getwd(),
EvalMethod = 'auc',
min_size = 4,
summarization = 'NetSHy',
BetweenShrinkage = 5,
ncomp_pls = 3)
Resulting output, tried from n_cols 5 to 400 to see if it was a sample issue
**********************************
* Welcome to Automated SmCCNet! *
**********************************
--------------------------------------------------
>> Starting data preprocessing...
--------------------------------------------------
This project uses multiomics CCA
--------------------------------------------------
>> Now determining the scaling factor for multi-omics analysis...
--------------------------------------------------
The scaling factor selection is:
rna-protein: 0.00534874045183213
rna-Phenotype: 1
protein-Phenotype: 1
--------------------------------------------------
>> Determining the best penalty selection through cross-validation...
--------------------------------------------------
Error in `purrr::map()`:
ℹ In index: 1.
Caused by error in `Y[foldIdx[[x]], ]`:
! subscript out of bounds
Run `rlang::last_trace()` to see where the error occurred.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels