Not sure if I'm missing something here or whether this is a bug. The paste generic defined in BiocGenerics seems to behave weirdly when 'lapply'd to a list of character vectors. A simple example follows:

Starting from a fresh R session:

library(BiocGenerics)
aa = list(letters[1:10], letters[11:20])
lapply(aa, paste, collapse = "")

The result is:

[[1]]
[1] "a" "b" "c" "d" "e" "f" "g" "h" "i" "j"

[[2]]
 [1] "k" "l" "m" "n" "o" "p" "q" "r" "s" "t"

The expected result is to collapse both vectors into a single string, as base::paste does:

> lapply(aa, base::paste, collapse = "")
[[1]]
[1] "abcdefghij"

[[2]]
[1] "klmnopqrst"

> sessionInfo()
R version 3.4.1 Patched (2017-08-11 r73090)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Fedora 26 (Twenty Six)

Matrix products: default
BLAS: /usr/local/atlas/lib/libtatlas.so
LAPACK: /usr/local/lib64/R-3.4.1-patched-atlas-parallel/lib/libRlapack.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

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

other attached packages:
[1] BiocGenerics_0.22.0

loaded via a namespace (and not attached):
[1] compiler_3.4.1

FWIW, in R-3.2.2 and its corresponding version of BioConductor this used to work fine, though I am not sure that paste was already part of BiocGenerics. Here's an older sessionInfo output under which paste behaved as expected (with BiocGenerics 0.16.1):

R version 3.2.2 (2015-08-14)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Fedora 24 (Twenty Four)

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C
 [9] LC_ADDRESS=C               LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C

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

other attached packages:
 [1] org.Mm.eg.db_3.2.3         DESeq2_1.10.1
 [3] RcppArmadillo_0.7.200.2.0  Rcpp_0.12.6
 [5] SummarizedExperiment_1.0.2 GenomicRanges_1.22.4
 [7] GenomeInfoDb_1.6.3         anRichment_0.82-1
 [9] WGCNA_1.60-90              fastcluster_1.1.20
[11] dynamicTreeCut_1.63-1      GO.db_3.2.2
[13] AnnotationDbi_1.32.3       IRanges_2.4.8
[15] S4Vectors_0.8.11           Biobase_2.30.0
[17] BiocGenerics_0.16.1        preprocessCore_1.32.0
[19] sva_3.18.0                 genefilter_1.52.1
[21] mgcv_1.8-13                nlme_3.1-128
[23] sqldf_0.4-10               RSQLite_1.0.0
[25] DBI_0.4-1                  gsubfn_0.6-6
[27] proto_0.3-10               matrixStats_0.50.2
[29] doParallel_1.0.10          iterators_1.0.8
[31] foreach_1.4.3              reshape_0.8.5
[33] Hmisc_3.17-4               ggplot2_2.1.0
[35] Formula_1.2-1              survival_2.39-5
[37] lattice_0.20-33            impute_1.44.0
[39] cluster_2.0.4              class_7.3-14
[41] MASS_7.3-45

loaded via a namespace (and not attached):
 [1] splines_3.2.2        latticeExtra_0.6-28  fit.models_0.5-14
 [4] robustbase_0.92-7    chron_2.3-47         RColorBrewer_1.1-2
 [7] XVector_0.10.0       colorspace_1.2-6     Matrix_1.2-6
[10] plyr_1.8.4           pcaPP_1.9-72         XML_3.98-1.4
[13] zlibbioc_1.16.0      xtable_1.8-2         mvtnorm_1.0-5
[16] scales_0.4.0         BiocParallel_1.4.3   annotate_1.48.0
[19] nnet_7.3-12          foreign_0.8-66       data.table_1.9.6
[22] locfit_1.5-9.1       munsell_0.4.3        lambda.r_1.1.9
[25] futile.logger_1.4.3  grid_3.2.2           robust_0.4-18
[28] tcltk_3.2.2          gtable_0.2.0         codetools_0.2-14
[31] rrcov_1.4-3          gridExtra_2.2.1      futile.options_1.0.0
[34] geneplotter_1.48.0   rpart_4.1-10         acepack_1.3-3.3
[37] DEoptimR_1.0-8