I am not sure what the right protocol is for reporting bugs in ExperimentHub, but here we go. It appears that the dataset ExperimentHub()[["EH359"]] (apparently a.k.a. curatedMetagenomicData::ZellerG_2014.marker_abundance) is an ExpressionSet whose exprs is a matrix of characters. The matrix can be converted to numeric, and all elements seem to represent legitimate numbers, but I wonder whether this should not be fixed upstream.

library("ExperimentHub")

eh = ExperimentHub()

# snapshotDate(): 2016-10-26
zeller = eh[["EH359"]]
# see ?curatedMetagenomicData and browseVignettes('curatedMetagenomicData') for documentation
# loading from cache ‘/Users/huber//.ExperimentHub/359’
str(exprs(zeller))
# chr [1:130272, 1:156] "1.8115942029" "17.0542635659" "55.5555555556" ...
# - attr(*, "dimnames")=List of 2
#  ..$ : chr [1:130272] "gi|333126069|ref|NZ_AEMJ01000490.1|:c656-105" #"gi|381149847|ref|NZ_JH604847.1|:635-1279" "gi|331001572|ref|NZ_GL883724.1|:311-544" "gi|381150020|ref|NZ_JH605020.1|:16763-17575" ...
#  ..$ : chr [1:156] "CCIS00146684ST-4-0" "CCIS00281083ST-3-0" "CCIS02124300ST-4-0" "CCIS02379307ST-4-0" ...
nonum = is.na(as.numeric(exprs(zeller)))
table(nonum)
#nonum
#  FALSE
#20322432

sessionInfo()
R version 3.3.1 (2016-06-21)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: OS X 10.12.1 (Sierra)

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

attached base packages:
[1] stats4    parallel  stats     graphics  grDevices utils     datasets
[8] methods   base     

other attached packages:
 [1] curatedMetagenomicData_1.0.0 phyloseq_1.18.0             
 [3] magrittr_1.5                 ExperimentHubData_1.0.0     
 [5] AnnotationHubData_1.4.0      futile.logger_1.4.3         
 [7] GenomicRanges_1.26.1         GenomeInfoDb_1.10.0         
 [9] IRanges_2.8.0                S4Vectors_0.12.0            
[11] Biobase_2.34.0               ExperimentHub_1.0.0         
[13] AnnotationHub_2.6.0          BiocGenerics_0.20.0         
[15] fortunes_1.5-3              

loaded via a namespace (and not attached):
 [1] httr_1.2.1                    splines_3.3.1                
 [3] jsonlite_1.1                  foreach_1.4.3                
 [5] shiny_0.14.2                  interactiveDisplayBase_1.12.0
 [7] RBGL_1.50.0                   Rsamtools_1.26.1             
 [9] RSQLite_1.0.0                 lattice_0.20-34              
[11] RUnit_0.4.31                  chron_2.3-47                 
[13] digest_0.6.10                 XVector_0.14.0               
[15] colorspace_1.2-7              htmltools_0.3.5              
[17] httpuv_1.3.3                  Matrix_1.2-7.1               
[19] plyr_1.8.4                    OrganismDbi_1.16.0           
[21] GEOquery_2.40.0               XML_3.98-1.4                 
[23] biomaRt_2.30.0                rBiopaxParser_2.14.0         
[25] zlibbioc_1.20.0               xtable_1.8-2                 
[27] scales_0.4.0                  getopt_1.20.0                
[29] optparse_1.3.2                BiocParallel_1.8.1           
[31] biocViews_1.42.0              mgcv_1.8-15                  
[33] ggplot2_2.1.0                 SummarizedExperiment_1.4.0   
[35] GenomicFeatures_1.26.0        survival_2.40-1              
[37] mime_0.5                      MASS_7.3-45                  
[39] nlme_3.1-128                  xml2_1.0.0                   
[41] vegan_2.4-1                   graph_1.52.0                 
[43] BiocInstaller_1.24.0          tools_3.3.1                  
[45] data.table_1.9.6              stringr_1.1.0                
[47] munsell_0.4.3                 cluster_2.0.5                
[49] AnnotationDbi_1.36.0          lambda.r_1.1.9               
[51] Biostrings_2.42.0             ade4_1.7-4                   
[53] rhdf5_2.18.0                  grid_3.3.1                   
[55] RCurl_1.95-4.8                iterators_1.0.8              
[57] biomformat_1.2.0              AnnotationForge_1.16.0       
[59] igraph_1.0.1                  bitops_1.0-6                 
[61] multtest_2.30.0               gtable_0.2.0                 
[63] codetools_0.2-15              DBI_0.5-1                    
[65] curl_2.2                      reshape2_1.4.2               
[67] R6_2.2.0                      GenomicAlignments_1.10.0     
[69] rtracklayer_1.34.1            futile.options_1.0.0         
[71] permute_0.9-4                 ape_3.5                      
[73] stringi_1.1.2                 Rcpp_0.12.7                  
[75] BiocCheck_1.10.0             

The issue is related to this function:

merged_eset <- function(assay_data, pheno_data, experiment_data) {
    as.matrix(assay_data) %>%
    ExpressionSet(., pheno_data, experimentData = experiment_data)
}

I have made the following change to ensure that all the matrices are numeric. The datasets will take just over 24 hours to regenerate and will be updated within ExperimentHub early next week. Sorry for the confusion.

merged_eset <- function(assay_data, pheno_data, experiment_data) {
    data.matrix(assay_data) %>%
    ExpressionSet(., pheno_data, experimentData = experiment_data)
}

Thanks Wolfgang - it's a pipeline bug that we'll fix and post here when that's done. 

Hello again, just wanted to give an update that the issue with the character matrix is fixed. All data in ExperimentHub should be correct! : )