Dear mailing list, I have a problem with rtraclayer that you can hopefully help me with. I'm trying to export tracks to the UCSC browser or to BED files through rtracklayer, but I have problems with the blockStarts and blockSizes fields functioning, and get a >export(targetTrack, format="bed", con="bed.bed") Error in df[[2]] + lastSize : non-numeric argument to binary operator error when trying to export a targetTrack with these columns. Removing the columns works fine. I then noticed this chunck of code in the source ( 1.10.4), in bed.R if (variant == "bedGraph") { if (is.null(score)) ## bedGraph requires score score <- 0 df$score <- score } else { blockSizes <- object$blockSizes blockStarts <- object$blockStarts if (variant == "bed15" && is.null(blockSizes)) blockStarts <- blockSizes <- "" # bed15 must have all 15 cols if (!is.null(blockSizes) || !is.null(blockStarts)) { if (is.null(blockSizes)) stop("'blockStarts' specified without 'blockSizes'") if (is.null(blockStarts)) stop("'blockSizes' specified without 'blockStarts'") * lastBlock <- function(x) sub(".*,", "", "20,20") lastSize <- lastBlock(blockSizes) lastStart <- lastBlock(blockStarts) if (any(df[[2]] + lastSize + lastStart != df[[3]]) ||* any(sub(",.*", "", blockStarts) != 0)) stop("blocks must span entire feature") blockCount <- length(strsplit(blockSizes, ",")[[1]]) } Notice the code in bold - it seems like the "20,20" is hardcoded, and doesn't take x as a parameter. Could this be a bug? Furthermore, wouldnt *lastSize* always be a character, rendering *if (any(df[[2]] + lastSize + lastStart != df[[3]]) *problematic? Best regards, and thx in advance, Johannes -- Johannes E. Waage, cand.scient. M.Sc.), Ph.D student Porse Group | Sandelin Group Biotech Research & Innovation Centre (BRIC) | Bioinformatics Centre, Dep. Of Bioligy University of Copenhagen Ole Maaløes Vej 5 DK-2200 Copenhagen N johannes.waage@bric.ku.dk <rune.damgaard@bric.ku.dk> | jwaage@binf.ku.dk http://www.bric.ku.dk | http://binf.ku.dk [[alternative HTML version deleted]]

Thanks for finding this. That's been broken for a year, but no one ran into it, I guess. Must have been distracted while writing the code. I've made the change for 1.10.5. Michael On Wed, Nov 3, 2010 at 4:34 AM, Johannes Waage <johannes.waage@bric.dk>wrote: > Dear mailing list, > > I have a problem with rtraclayer that you can hopefully help me with. > I'm trying to export tracks to the UCSC browser or to BED files through > rtracklayer, but I have problems with the blockStarts and blockSizes fields > functioning, and get a > > >export(targetTrack, format="bed", con="bed.bed") > Error in df[[2]] + lastSize : non-numeric argument to binary operator > > error when trying to export a targetTrack with these columns. Removing the > columns works fine. > > I then noticed this chunck of code in the source ( 1.10.4), in bed.R > > if (variant == "bedGraph") { > if (is.null(score)) ## bedGraph requires score > score <- 0 > df$score <- score > } else { > blockSizes <- object$blockSizes > blockStarts <- object$blockStarts > if (variant == "bed15" && is.null(blockSizes)) > blockStarts <- blockSizes <- "" # bed15 must have all 15 > cols > if (!is.null(blockSizes) || !is.null(blockStarts)) { > if (is.null(blockSizes)) > stop("'blockStarts' specified without 'blockSizes'") > if (is.null(blockStarts)) > stop("'blockSizes' specified without 'blockStarts'") > * lastBlock <- function(x) sub(".*,", "", "20,20") > lastSize <- lastBlock(blockSizes) > lastStart <- lastBlock(blockStarts) > if (any(df[[2]] + lastSize + lastStart != df[[3]]) ||* > any(sub(",.*", "", blockStarts) != 0)) > stop("blocks must span entire feature") > blockCount <- length(strsplit(blockSizes, ",")[[1]]) > } > > Notice the code in bold - it seems like the "20,20" is hardcoded, and > doesn't take x as a parameter. Could this be a bug? > Furthermore, wouldnt *lastSize* always be a character, rendering *if > (any(df[[2]] + lastSize + lastStart != df[[3]]) *problematic? > > Best regards, and thx in advance, > > Johannes > > > -- > Johannes E. Waage, cand.scient. M.Sc.), > Ph.D student > Porse Group | Sandelin Group > > Biotech Research & Innovation Centre (BRIC) | Bioinformatics Centre, Dep. > Of > Bioligy > University of Copenhagen > Ole Maaløes Vej 5 > DK-2200 Copenhagen N > > johannes.waage@bric.ku.dk <rune.damgaard@bric.ku.dk> | jwaage@binf.ku.dk > http://www.bric.ku.dk | http://binf.ku.dk > > [[alternative HTML version deleted]] > > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > [[alternative HTML version deleted]]