Recently, I had a problem of classifying samples into subtypes based on several gene signatures.

After reading and testing the source code of ImmuneSubtypeClassifier, I think the basis of ImmuneSubtypeClassifier——Top Scoring Pairs, could be generalized to any new gene signatures, which is my demand.

I had built a new R package (not being pushed onto Github yet) that could classify samples into subtypes based on any gene signatures that users give (including the gene signatures used in ImmuneSubtypeClassifier package).

I plan to publish my paper in the future, and the new package would be part of it. But I‘m not sure whether it is proper to publish it directly, because your source code about Top Scoring Pairs algorithm was deeply referred to.

Would you give some suggestions? Thanks a lot!

Hi,
a very good tool helps us classify tumors into different immune subtypes!
I run this function and get the results. in the results table, I found there are some discordances between BestCall and probabilities of belonging to each subtype. For instance, in this sample, bestcall indicates C3 subclass but C4 has higher probability. So I don't know which result I need to keep?

and here is the summary infomation in the series. 18 samples have these inconsistent information

Thanks in advance!

Hi Dave, I'm using this commit: 30e6215

and getting this error:

calls <- ImmuneSubtypeClassifier::callEnsemble(
input_matrix,
numCores = args$num_cores
)

Error in ImmuneSubtypeClassifier::callEnsemble(input_matrix, numCores = args$num_cores) :
unused argument (numCores = args$num_cores)
Execution halted

Here is the error message.

Installing package into ‘/home/xxxxxxx/Rlib/R_4.0’
(as ‘lib’ is unspecified)

• installing source package ‘ImmuneSubtypeClassifier’ ...
  ** using staged installation
  ** R
  ** data
  *** moving datasets to lazyload DB
  ** inst
  ** byte-compile and prepare package for lazy loading
  Error: (converted from warning) package ‘ggplot2’ was built under R version 4.0.3
  Execution halted
  ERROR: lazy loading failed for package ‘ImmuneSubtypeClassifier’
• removing ‘/home/xxxxxxx/Rlib/R_4.0/ImmuneSubtypeClassifier’
  Error: Failed to install 'ImmuneSubtypeClassifier' from GitHub:
  (converted from warning) installation of package ‘/tmp/Rtmp3GLvxl/file18c167b252f2/ImmuneSubtypeClassifier_0.1.0.tar.gz’ had non-zero exit status

Hi,

This is a great package that can help our research a lot, even though a little bit difficult to install. However, these are not the emphasis.

I have used this package for several datasets classification, however, I found that only in TCGA datasets, it can perform well. for several GEO dataset (I have tried 7 sets, all are microarray), I can get C2 subtype only.

Then I tried a TARGET dataset, which is sequenced by RNA-seq. I can get C1, C4 and only a few C6/C2 subtypes. I don't know why? Maybe this dataset do have only a few C6/C2 subtypes.

So, I wonder whether this package can not be used for microarray data? but it's built by pearson correlation analysis, I do not think microarray data is not applicable.

sincerely yours!

Hi

the ImmuneSubtypeClassifier is very useful, but i want to know the accuracy of model in test data. i don't find it, so can you share with me?

                                                                                                                                             ningwei

Hi Dave, I'm using this commit: 30e6215

and getting this error:

calls <- ImmuneSubtypeClassifier::callEnsemble(input_matrix)

Error in geneMatch(X, geneids) : object 'ebpp_genes_sig' not found
Calls: -> geneMatch
In addition: Warning messages:
1: In data("subtype_caller_model") :
data set ‘subtype_caller_model’ not found
2: In data("ensemble_model") : data set ‘ensemble_model’ not found
3: In data(ebpp_gene) : data set ‘ebpp_gene’ not found
Execution halted

With the demo data Xmat, the callEnsembl function works fine but with COAD data (subset of 20 samples) I get this error. It does not matter if symbols or entrez IDs are used. A histogram of the values for Xmat and the COAD data are almost identical (attached). I've dug thru the code and its failing in the dataProc function inside the callOneSubtype function. I'll continue down the chain of functions but I'm already 3 levels deep. I also tried installing the older xgboost as recommended but that didn't resolve the issue.

Error in rownames<-(*tmp*, value = gs) :
attempt to set 'rownames' on an object with no dimensions

Below is my sessionInfo (with xgboost_1.5

sessionInfo()
R version 4.1.1 (2021-08-10)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Big Sur 11.6.1

Matrix products: default
LAPACK: /Library/Frameworks/R.framework/Versions/4.1/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] parallel stats graphics grDevices utils datasets methods base

other attached packages:
[1] RColorBrewer_1.1-2 ImmuneSubtypeClassifier_0.1.0 plotROC_2.2.1
[4] xgboost_1.5.0.2 gridExtra_2.3 forcats_0.5.1
[7] stringr_1.4.0 purrr_0.3.4 readr_2.1.1
[10] tidyr_1.1.4 tibble_3.1.6 tidyverse_1.3.1
[13] dplyr_1.0.7 ggplot2_3.3.5

loaded via a namespace (and not attached):
[1] Rcpp_1.0.7 lubridate_1.8.0 lattice_0.20-45 assertthat_0.2.1 digest_0.6.29 utf8_1.2.2
[7] R6_2.5.1 cellranger_1.1.0 backports_1.4.0 reprex_2.0.1 evaluate_0.14 httr_1.4.2
[13] pillar_1.6.4 rlang_0.4.12 readxl_1.3.1 rstudioapi_0.13 data.table_1.14.2 jquerylib_0.1.4
[19] Matrix_1.3-4 rmarkdown_2.11 bit_4.0.4 munsell_0.5.0 broom_0.7.10 compiler_4.1.1
[25] modelr_0.1.8 xfun_0.28 pkgconfig_2.0.3 htmltools_0.5.2 tidyselect_1.1.1 fansi_0.5.0
[31] crayon_1.4.2 tzdb_0.2.0 dbplyr_2.1.1 withr_2.4.3 grid_4.1.1 jsonlite_1.7.2
[37] gtable_0.3.0 lifecycle_1.0.1 DBI_1.1.1 magrittr_2.0.1 scales_1.1.1 vroom_1.5.7
[43] cli_3.1.0 stringi_1.7.6 fs_1.5.2 bslib_0.3.1 xml2_1.3.3 ellipsis_0.3.2
[49] generics_0.1.1 vctrs_0.3.8 tools_4.1.1 bit64_4.0.5 glue_1.5.1 hms_1.1.1
[55] fastmap_1.1.0 yaml_2.2.1 colorspace_2.0-2 rvest_1.0.2 knitr_1.36 haven_2.4.3
[61] sass_0.4.0

Hi,

The package seems to be incompatible with the last version of xgboost, probably because of the issue outlined here -
dmlc/xgboost#5794

Using version 1.0.0.2 of xgboost works OK (albeit it warns that models were built in xgboost < 1.0.0, might be good to upgrade them anyway).

The error message I got was -
Error in predict.xgb.Booster(mi$bst, Xbin) :
[12:28:19] amalgamation/../src/learner.cc:506: Check failed: mparam_.num_feature != 0 (0 vs. 0) : 0 feature is supplied. Are you using raw Booster interface?

Hi,
it is very useful package! I download fpkm gene expression data from TCGA by TCGAbiolinks package, and i try to get the immune subtypes of the LUAD samples by using this package.when i compare this subtypes result with the subtypes of this paper(five_signature_mclust_ensemble_results.tsv.gz),their subtypes are different.
But when i use the gene expression of paper(ebppSubset.tsv.bz2) to classify these patients,the subtype result can match with the paper result(five_signature_mclust_ensemble_results.tsv.gz).So i want to know if there is any problem to classify immune subtype by using fpkm gene expression data from TCGA by TCGAbiolinks.if i need to do any thing with the gene expression download from TCGA?

Thanks!
here are some codes and results!
code and result .pptx

When I run the classifier with one sample,
I get error below:

Error in `rownames<-`(`*tmp*`, value = ebpp_genes_sig$Symbol) : 
  attempt to set 'rownames' on an object with no dimensions
Calls: capture.output ... withCallingHandlers -> <Anonymous> -> geneMatch -> rownames<-
Execution halted

With two samples, it works smoothly.

Hi,

I used a subset of data from TCGA for test. However, I met the error like this:

Error in predict.xgb.Booster(mi$bst, Xbin) : 
  [22:12:42] amalgamation/../src/learner.cc:506: Check failed: mparam_.num_feature != 0 (0 vs. 0) : 0 feature is supplied.  Are you using raw Booster interface? 

Could you help to solve this?
I think it might be good to add a test data for others to use this tool !

Hello~
After some tests about this package, I think it's easy to use. Meanwhile, the capacity of concordance is robust, according to the reported result.Here are my questions:

1.Are there any limitations about the form of gene expression data? For example,is it available if I use gene expression data from Affymetrix miroarray? Or it's only fitted in RPKM data?

2.Before go to the pipeline, shall gene matrix be normalized or scaled?
Thanks~

Hello,

Firstly, thank you for compiling this resource for the public. The whole work is an incredible inspiration to me and others. I have been using gene expression data and exploring these immune subtypes. I noticed that there is a significant difference in best immune subtype calls depending on whether I run RSEM expected counts vs TPM gene expression data. The TPM data results in a much higher proportion of C4 in my data. I was wondering if in theory, one or the other would be more "valid"?

Thanks for your time

many thanks for providing this helpful package. Can you tell me what the input values should be for the matrix. Raw counts? TPM etc? I see from the previous version you say it is sensitive to this. Would be grateful for advice