parallel run (nCores > 1) always fails

[slowsmile]

Dear JunctionSeq author
I am running the JunctionSeq on both the test dataset that comes with the package and my own dataset, the function runs well with the default option of nCores = 1 , however when I specified nCores > 1, the function "runJunctionSeqAnalyses" always fails with the error reported as length of 'dimnames' [2] not equal to array extent

my code of running this function on test dataset is below

require("JunctionSeq")

decoder.file <- system.file("extdata/annoFiles/decoder.bySample.txt", package="JctSeqData", mustWork=TRUE);
decoder <- read.table(decoder.file, header = TRUE, stringsAsFactors = FALSE);
gff.file <- system.file("extdata/cts/withNovel.forJunctionSeq.gff.gz", package = "JctSeqData", mustWork = TRUE);

countFiles <- system.file(paste0("extdata/cts/", decoder$sample.ID, "/QC.spliceJunctionAndExonCounts.withNovel.forJunctionSeq.txt.gz"),
                          package = "JctSeqData", mustWork = TRUE);

jscs <- runJunctionSeqAnalyses(sample.files = countFiles, sample.names = decoder$sample.ID, condition=factor(decoder$group.ID),
                               flat.gff.file = gff.file, nCores = 3, analysis.type = "junctionsAndExons");

then I got the following error

 STARTING runJunctionSeqAnalyses (v1.5.4) (Thu Dec 13 16:47:11 2018)
> rJSA: sample.files:  /home/xxu1/R/x86_64-pc-linux-gnu-library/3.5/JctSeqData/extdata/cts/SAMP2/QC.spliceJunctionAndExonCounts.withNovel.forJunctionSeq.txt.gz, /home/xxu1/R/x86_64-pc-linux-gnu-library/3.5/JctSeqData/extdata/cts/SAMP3/QC.spliceJunctionAndExonCounts.withNovel.forJunctionSeq.txt.gz, /home/xxu1/R/x86_64-pc-linux-gnu-library/3.5/JctSeqData/extdata/cts/SAMP4/QC.spliceJunctionAndExonCounts.withNovel.forJunctionSeq.txt.gz, /home/xxu1/R/x86_64-pc-linux-gnu-library/3.5/JctSeqData/extdata/cts/SAMP5/QC.spliceJunctionAndExonCounts.withNovel.forJunctionSeq.txt.gz, /home/xxu1/R/x86_64-pc-linux-gnu-library/3.5/JctSeqData/extdata/cts/SAMP6/QC.spliceJunctionAndExonCounts.withNovel.forJunctionSeq.txt.gz
> rJSA: sample.names:  SAMP1, SAMP2, SAMP3, SAMP4, SAMP5, SAMP6
> rJSA: condition:  CASE, CASE, CASE, CTRL, CTRL, CTRL
> rJSA: analysis.type:  junctionsAndExons
> rJSA: use.junctions:  TRUE
> rJSA: use.novel.junctions:  TRUE
> rJSA: use.exons:  TRUE
> rJSA: nCores:  3
> rJSA: use.covars:
> rJSA: test.formula0:  ~ sample + countbin
> rJSA: test.formula1:  ~ sample + countbin + condition:countbin
> rJSA: use.multigene.aggregates:  FALSE
> rJSA: Reading Count files... Thu Dec 13 16:47:11 2018.
-> STARTING readJunctionSeqCounts (Thu Dec 13 16:47:11 2018)
---> RJSC; (v1.5.4)
---> RJSC: samplenames: SAMP1,SAMP2,SAMP3,SAMP4,SAMP5,SAMP6
---> RJSC: flat.gff.file: /home/xxu1/R/x86_64-pc-linux-gnu-library/3.5/JctSeqData/extdata/cts/withNovel.forJunctionSeq.gff.gz
---> RJSC: use.exons:TRUE
---> RJSC: use.junctions:TRUE
---> RJSC: use.novel.junctions:TRUE
---> File read complete.
---> Extracted counts. Found 16167 features so far.
Error in dimnames(x) <- dn :
  length of 'dimnames' [2] not equal to array extent

below is my sessionInfo()

R version 3.5.0 (2018-04-23) Platform: x86_64-pc-linux-gnu (64-bit) Running under: Scientific Linux 7.4 (Nitrogen) Matrix products: default BLAS/LAPACK: /apps/prod/easybuild/sl7.x86_64/software/OpenBLAS/0.2.20-GCC-6.4.0-2.28/lib/libopenblasp-r0.2.20.so locale: [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C 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 LC_PAPER=en_US.UTF-8 LC_NAME=C [9] LC_ADDRESS=C LC_TELEPHONE=C LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C attached base packages: [1] parallel stats4 stats graphics grDevices utils datasets methods base other attached packages: [1] JunctionSeq_1.10.0 RcppArmadillo_0.9.200.5.0 Rcpp_1.0.0 SummarizedExperiment_1.10.1 [5] DelayedArray_0.6.0 BiocParallel_1.14.2 matrixStats_0.54.0 Biobase_2.40.0 [9] GenomicRanges_1.32.7 GenomeInfoDb_1.16.0 IRanges_2.14.12 S4Vectors_0.18.3 [13] BiocGenerics_0.26.0 loaded via a namespace (and not attached): [1] bit64_0.9-7 splines_3.5.0 Formula_1.2-3 assertthat_0.2.0 statmod_1.4.30 [6] latticeExtra_0.6-28 blob_1.1.1 GenomeInfoDbData_1.1.0 yaml_2.1.19 pillar_1.2.2 [11] RSQLite_2.1.1 backports_1.1.2 lattice_0.20-35 glue_1.3.0 digest_0.6.18 [16] RColorBrewer_1.1-2 XVector_0.20.0 checkmate_1.8.5 colorspace_1.3-2 htmltools_0.3.6 [21] Matrix_1.2-15 plyr_1.8.4 DESeq2_1.20.0 XML_3.98-1.16 pkgconfig_2.0.2 [26] genefilter_1.62.0 zlibbioc_1.26.0 purrr_0.2.5 xtable_1.8-3 scales_1.0.0 [31] htmlTable_1.11.2 tibble_1.4.2 annotate_1.58.0 ggplot2_3.1.0 nnet_7.3-12 [36] lazyeval_0.2.1 survival_2.42-3 magrittr_1.5 memoise_1.1.0 foreign_0.8-70 [41] tools_3.5.0 data.table_1.11.8 stringr_1.3.1 locfit_1.5-9.1 munsell_0.5.0 [46] cluster_2.0.7-1 plotrix_3.7-1 AnnotationDbi_1.42.1 bindrcpp_0.2.2 compiler_3.5.0 [51] rlang_0.3.0.1 grid_3.5.0 RCurl_1.95-4.11 rstudioapi_0.8 htmlwidgets_1.2 [56] bitops_1.0-6 base64enc_0.1-3 gtable_0.2.0 DBI_1.0.0 R6_2.3.0 [61] gridExtra_2.3 knitr_1.20 dplyr_0.7.7 bit_1.1-13 bindr_0.1.1 [66] Hmisc_4.1-1 stringi_1.2.2 geneplotter_1.58.0 rpart_4.1-13 acepack_1.4.1 [71] tidyselect_0.2.5

Any suggestion? Thanks

[hartleys]

Multiple ncores is currently deprecated. During JunctionSeq's development the bioconductor parallellization framework changed several times without support for backwards compatibility. So basically nCores only works on certain specific versions of R.

Just use nCores = 1. The parallelization didn't seem to help that much anyways.