Hi Dr. Haibe-Kains,

I'm using the package MetaGxBreast to extract gene expression data from METABRIC and TCGA. However, it seems the function was not working recently for MetaGxBreast (error message attached).

Would you please help to solve this issue?

Thank you!

library(MetaGxBreast)
esets = MetaGxBreast::loadBreastEsets(loadString = c("TCGA","METABRIC"))[[1]]
snapshotDate(): 2021-10-19
see ?MetaGxBreast and browseVignettes('MetaGxBreast') for documentation
loading from cache
see ?MetaGxBreast and browseVignettes('MetaGxBreast') for documentation
loading from cache
Error: failed to load resource
name: EH1099
title: METABRIC
reason: error in evaluating the argument 'x' in selecting a method for function 'get': bad restore file magic number (file may be corrupted) -- no data loaded
In addition: Warning message:
file ‘4e9e48f5de98_1099’ has magic number 'SQLit'
Use of save versions prior to 2 is deprecated

Dear all,
I got those errors when trying to load the TCGA data set:
esets = MetaGxBreast::loadBreastEsets(loadString = c('tcga'))[[1]]
snapshotDate(): 2020-10-27
Error in MetaGxBreast::loadBreastEsets(loadString = c("tcga")) :
loadString needs to be one of majority, metabric, tcga, or a character vector of datasets to load from the hub

and a variant of the same problem

esets = MetaGxBreast::loadBreastEsets(loadString = c("CAL", "tcga"))[[1]]
snapshotDate(): 2020-10-27
Error in MetaGxBreast::loadBreastEsets(loadString = c("CAL", "tcga")) :
tcga could not be found in the MetaGxBreast package in the experiment hub

could you help ?
thanks
Ataaillah

I am attempting to use all data sets for which the outcome is survival rather than categorical. loadBreastEsets(minNumberEvents = 1) results in list(), an empty list. But, looking at the vignette, I see at least some data sets do have survival outcome variables such as vital_status and recurrence_status in the clinical variable heatmap plot on the last page. Is it a bug?

RFS data (t.rfs and e.rfs) for dataset UPP in the DFS mRNA compendium shared by the Sotirou lab are identical to disease-specific survival (DSS) in the original data available at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE3494

• DSS TIME (Disease-Specific Survival Time in years)
• DSS EVENT (Disease-Specific Survival EVENT; 1=death from breast cancer, 0=alive or censored)

[Reported by Francois Richard and Theodoros Foukakis' team]

What valid values can a user can specify? This is not documented anywhere that I can see. I tried two varieties without success.

hub = ExperimentHub::ExperimentHub()
breastData <- query(hub, c("MetaGxBreast", "ExpressionSet"))
> breastData
  ExperimentHub with 39 records
             title 
  EH1076 | CAL   
  EH1077 | DFHCC 
  EH1078 | DFHCC2
  EH1079 | DFHCC3
  EH1080 | DUKE  
  ...      ...   
  EH1110 | UCSF  
  EH1111 | UNC4
  
> x <- loadBreastEsets("EH1080")
  Error in loadBreastEsets("EH1080") : loadString needs to be one of majority, metabric, tcga, or a character vector of
    datasets to load from the hub
> x <- loadBreastEsets("DUKE")
  Error in loadBreastEsets("DUKE") : loadString needs to be one of majority, metabric, tcga, or a character vector of
    datasets to load from the hub 

Bizarrely, x <- loadBreastEsets(c("DUKE", "DUKE2")) works. Is loadBreastEsets struggling with vectors of length one?

after loading TCGA data, I found that row.names(phenotype) is different from the columns "sample_name" "alt_sample_name".
the row.names(phenotype) is identical with colnames(expr).
I tend to believe that the programmer initially got error message since row.names(phenotype) != colnames(expr), and then forced them to be equal by assigning:
row.names(phenotype) = colnames(expr)

Then, this is only to avoid error message when creating eset object, and row.names(phenotype) is incorrect and we should ignore.

An I correct?

   Sorry for bother you.
   Recently, my friend and I have encountered an error when querying and downloading the public datasets curated in MetaGxBreast R package (as shown in attached screeshot). It seems that the public datasets with expression profiles and clinical data become unaccessible.
   We are wondering wether it is possible to solve this kind of problem.
   We are looking forward for your reply.
   Best wishes.

![WX20231009-002928@2x](https://github.com/bhklab/MetaGxBreast/assets/133837857/643b312e-2475-4aec-b9b0-2d1069767fa

e)

Thank you for creating this valuable resource.

I am interested in examining the association of a specific gene with survival. For this, I want to perform a survival analysis of multiple datasets together.

I assume I can combine expression data of the gene from the multiple sets without any need for post-processing like batch-effect adjustment. Please correct me if I am wrong.

How do I go about putting together the gene expression and meta (clinical) data of the multiple datasets? Is there a simple command to do this?

Survival data for dataset STK in the DFS mRNA compendium shared by the Sotirou lab is a very early and unfortunately slightly incorrect version that was initially deposited at GEO. Correct survival data can be found at
https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE1456

[Reported by Francois Richard and Theodoros Foukakis' team]

imputeMissing: remove patients from datasets with missing expression values
Does it impute missing values? Or, does it remove samples with them? These two are opposite actions.

dupplicates a list containing the names of patients that are believed to be dulicates across datasets (Sp Duplicates)

loadBreastDatasets Function to load breast cancer SummarizedExperiment objectsfrom the Experiment Hub (Sp objectsfrom)

Vignettes Page2

Datasets: Retain studies with a minimum umber of genes (default = 0) (Sp umber -> Number)

How can we get TCGA and METABRIC datasets using MetaGxBreast?