SummarizedExperiment Inheritance issue

[arcolombo]

The S4 structure is gathering a list and not assembling a proper s4 experiment. works on my linux, not on mac. maybe class conflict with SummarizedExperiment, or GenomicRanges.

n object of class "KallistoExperiment" Slot "transcriptomes": [1] "ErccDbLite.ERCC.97, EnsDbLite.Hsapiens.81, RepDbLite.Hsapiens.2007"

Slot "kallistoVersion": [1] "0.42.3"

Slot "rowRanges": GRanges object with 213782 ranges and 9 metadata columns: seqnames ranges strand | tx_length gc_content tx_id

| ERCC-00002 Unknown [1, 1] * | 1 ERCC-00003 Unknown [1, 1] * | 1 ERCC-00004 Unknown [1, 1] * | 1 ERCC-00007 Unknown [1, 1] * | 1 ERCC-00009 Unknown [1, 1] * | 1 ... ... ... ... . ... ... ... SVA_E Unknown [1, 1] * | 1 SVA_F Unknown [1, 1] * | 1 AluYb11 Unknown [1, 1] * | 1 AluYb10 Unknown [1, 1] * | 1 AluYb8a1 Unknown [1, 1] * | 1 gene_id gene_name entrezid tx_biotype gene_biotype ERCC-00002 ERCC-00003 ERCC-00004 ERCC-00007 ERCC-00009 ... ... ... ... ... ... SVA_E SVA_F AluYb11 AluYb10 AluYb8a1 biotype_class ERCC-00002 ERCC-00003 ERCC-00004 ERCC-00007 ERCC-00009 ... ... SVA_E SVA_F AluYb11 AluYb10 AluYb8a1 ------- seqinfo: 1 sequence from an unspecified genome Slot "colData": DataFrame with 6 rows and 1 column ID n1 n1 n2 n2 n4 n4 s1 s1 s2 s2 s4 s4 Slot "assays": Reference class object of class "ShallowSimpleListAssays" Field "data": List of length 3 names(3): est_counts eff_length tpm Slot "NAMES": NULL Slot "elementMetadata": DataFrame with 213782 rows and 0 columns Slot "metadata":

[arcolombo]

this issue is on a mac. there is an issue with the KallistoExperiment constructor

on my system did not have the error sessionInfo() R version 3.4.0 (2017-04-21) Platform: x86_64-pc-linux-gnu (64-bit) Running under: Ubuntu 14.04.5 LTS

Matrix products: default BLAS: /usr/lib/libblas/libblas.so.3.0 LAPACK: /usr/lib/lapack/liblapack.so.3.0

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] stats4 parallel stats graphics grDevices utils datasets [8] methods base

other attached packages: [1] SummarizedExperiment_1.0.2 Biobase_2.30.0
[3] GenomicRanges_1.22.4 GenomeInfoDb_1.6.3
[5] IRanges_2.4.8 S4Vectors_0.8.11
[7] BiocGenerics_0.16.1 BiocInstaller_1.20.3

loaded via a namespace (and not attached): [1] zlibbioc_1.16.0 compiler_3.4.0 XVector_0.10.0 tools_3.4.0

[arcolombo]

arkas is not compatible with SummarizedExperiment 1.6.1 (bioCLite version 3.5)

[arcolombo]

so this is a bit odd.

first, the issue is with the new release SummarizedExperiment (1.6.1), Bioconductor release inheritance with the arkas class "KallistoExperiment".

the problem is that when you construct a "KallistoExperiment" using the SummarizedExperiment class, adding the kallisto version, and transcriptomes, it returns a list of all the slots, and the S4 object is not properly assembled.

If I change the class name from "KallistoExperiment" to "KExperiment", with the exact same constructor (see below)

setClass("KExperiment",representation(transcriptomes="character",kallistoVersion="character"),contains="RangedSummarizedExperiment")

new("KExperiment",Y,kallistoVersion="50",transcriptomes="MM") class: KExperiment dim: 213782 6 metadata(0): assays(3): est_counts eff_length tpm rownames(213782): ERCC-00002 ERCC-00003 ... AluYb10 AluYb8a1 rowData names(9): tx_length gc_content ... gene_biotype biotype_class colnames(6): n1 n2 ... s2 s4 colData names(1): ID

new("KExperiment",Y,kallistoVersion="50",transcriptomes="MM")->p Browse[2]> head(str(p)) Formal class 'KExperiment' [package ".GlobalEnv"] with 8 slots ..@ transcriptomes : chr "MM" ..@ kallistoVersion: chr "50" ..@ rowRanges :Formal class 'GRanges' [package "GenomicRanges"] with 6 slots .. .. ..@ seqnames :Formal class 'Rl

here we see the two character slots from the "KExperiment" definition. note that the object Y in the above creation is the SummarizedExperiment of the Normal Senescent arkasData.

is "KallistoExperiment" a reserved word for BioConductor Devel?

[ttriche]

Looking into this. I thought so at first, but a quick grep through SummarizedExperiment suggests otherwise. Hmm.

--t

On Wed, May 24, 2017 at 8:55 AM, Anthony R. Colombo < notifications@github.com> wrote:

so this is a bit odd.

first, the issue is with the new release SummarizedExperiment (1.6.1), Bioconductor release inheritance with the arkas class "KallistoExperiment".

the problem is that when you construct a "KallistoExperiment" using the SummarizedExperiment class, adding the kallisto version, and transcriptomes, it returns a list of all the slots, and the S4 object is not properly assembled.

If I change the class name from "KallistoExperiment" to "KExperiment", with the exact same constructor (see below)

setClass("KExperiment",representation(transcriptomes= "character",kallistoVersion="character"),contains=" RangedSummarizedExperiment")

new("KExperiment",Y,kallistoVersion="50",transcriptomes="MM") class: KExperiment dim: 213782 6 metadata(0): assays(3): est_counts eff_length tpm rownames(213782): ERCC-00002 ERCC-00003 ... AluYb10 AluYb8a1 rowData names(9): tx_length gc_content ... gene_biotype biotype_class colnames(6): n1 n2 ... s2 s4 colData names(1): ID

new("KExperiment",Y,kallistoVersion="50",transcriptomes="MM")->p Browse[2]> head(str(p)) Formal class 'KExperiment' [package ".GlobalEnv"] with 8 slots ..@ transcriptomes : chr "MM" ..@ kallistoVersion: chr "50" ..@ rowRanges :Formal class 'GRanges' [package "GenomicRanges"] with 6 slots .. .. ..@ seqnames :Formal class 'Rl

here we see the two character slots from the "KExperiment" definition. note that the object Y in the above creation is the SummarizedExperiment of the Normal Senescent arkasData.

is "KallistoExperiment" a reserved word for BioConductor Devel?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/RamsinghLab/arkas/issues/45#issuecomment-303768468, or mute the thread https://github.com/notifications/unsubscribe-auth/AAARIp8_dtYmbsTWO0_p9Odj7btIOXiLks5r9FLzgaJpZM4NkZqI .

[arcolombo]

OK. I'm dropping the alias to see if that is an issue.

[arcolombo]

not any naming specific issue.

[arcolombo]

the problem lies within the Kallisto-methods.R

for instance, if I remove the methods, and install arkas without the Kallisto-methods.R the kexp is merged correctly.

suppressPackageStartupMessages(library(TxDbLite)) samples<-c("n1","n2","n4","s1","s2","s4") pathBase<-system.file("extdata",package="arkasData") merged <- mergeKallisto(samples, outputPath=pathBase) Bootstraps ignored for /Library/Frameworks/R.framework/Versions/3.4/Resources/library/arkasData/extdata/n1, set summarize=TRUE to use Bootstraps ignored for /Library/Frameworks/R.framework/Versions/3.4/Resources/library/arkasData/extdata/n2, set summarize=TRUE to use Bootstraps ignored for /Library/Frameworks/R.framework/Versions/3.4/Resources/library/arkasData/extdata/n4, set summarize=TRUE to use Bootstraps ignored for /Library/Frameworks/R.framework/Versions/3.4/Resources/library/arkasData/extdata/s1, set summarize=TRUE to use Bootstraps ignored for /Library/Frameworks/R.framework/Versions/3.4/Resources/library/arkasData/extdata/s2, set summarize=TRUE to use Bootstraps ignored for /Library/Frameworks/R.framework/Versions/3.4/Resources/library/arkasData/extdata/s4, set summarize=TRUE to use Setting transcriptome automatically from Kallisto call string. merged class: KallistoExperiment dim: 213782 6 metadata(0): assays(3): est_counts eff_length tpm rownames(213782): ERCC-00002 ERCC-00003 ... AluYb10 AluYb8a1 rowData names(9): tx_length gc_content ... gene_biotype biotype_class colnames(6): n1 n2 ... s2 s4 colData names(1): ID

this implies that the problem is not with the class, constructor, or the merge command. but some method. my first guess is to drop the alias commands. updates soon.

[arcolombo]

so the conflict has to do with this method

' @export

setAs("KallistoExperiment", "SummarizedExperiment", function(from) { metanames <- names(from@metadata) metaorder <- c("transcriptomes", "kallistoVersion", metanames) from@metadata$transcriptomes <- from@transcriptomes from@metadata$kallistoVersion <- from@kallistoVersion from@metadata <- from@metadata[metaorder] if (!identical(colnames(from@assays$data[[1]]),rownames(from@colData))){ for (i in names(from@assays$data)) { colnames(from@assays$data[[i]]) <- rownames(from@colData) } } SummarizedExperiment(assays=from@assays$data, rowRanges=from@rowRanges, colData=from@colData, metadata=from@metadata) })

how do I know this? I deleted the methods and added each method individually until the error popped up, and lo-and -behold, the merging broke upon the addition of the first setAs.

should we remove this or patch?

[arcolombo]

okay so the fix is to just remove this method from Kallisto-methods.R. perhaps one work around would be to write an R script to invoke it directly and simply move this method out of Kallisto-method.R and move this into the R directory. there is some implicit issue here, which is a mystery to me on why this error happens. my suggestion would be to create a function KEtoSE.R and copy this there.

[arcolombo]

this issue is resolved per my branch. PR forthcoming. created funcitons SEtoKE etc, and upgraded my end... worked... the vignettes built.... closing...