error with TackNames and UCSC

[eandresleon]

More info at:

https://support.bioconductor.org/p/103829/

[lawremi]

Already fixed.

[eandresleon]

Thanks. I've tested in 1.38.2 and it worked. Thanks again

[marencc]

I've teste 1.38.2 and still get the same error.

> refSeq             <- makeTxDbFromUCSC(genom="hg19",tablename="refGene")
Error in names(trackIds) <- sub("^ ", "", sapply(nodes, xmlValue)) : 
  'names' attribute [210] must be the same length as the vector [208]

> sessionInfo() R version 3.4.0 (2017-04-21) Platform: x86_64-pc-linux-gnu (64-bit) Running under: Ubuntu 16.04.2 LTS Matrix products: default BLAS: /usr/lib/libblas/libblas.so.3.6.0 LAPACK: /usr/lib/lapack/liblapack.so.3.6.0 locale: [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C LC_TIME=es_ES.UTF-8 LC_COLLATE=en_US.UTF-8 LC_MONETARY=es_ES.UTF-8 [6] LC_MESSAGES=en_US.UTF-8 LC_PAPER=es_ES.UTF-8 LC_NAME=C LC_ADDRESS=C LC_TELEPHONE=C [11] LC_MEASUREMENT=es_ES.UTF-8 LC_IDENTIFICATION=C attached base packages: [1] parallel stats4 stats graphics grDevices utils datasets methods base other attached packages: [1] BiocInstaller_1.28.0 rtracklayer_1.38.2 magrittr_1.5 forcats_0.2.0 stringr_1.2.0 dplyr_0.7.4 [7] purrr_0.2.4 readr_1.1.1 tibble_1.3.4 ggplot2_2.2.1 tidyverse_1.2.1 broom_0.4.3 [13] reshape2_1.4.3 annotate_1.56.1 XML_3.98-1.9 org.Hs.eg.db_3.5.0 GenomicFeatures_1.30.0 AnnotationDbi_1.40.0 [19] Biobase_2.38.0 tidyr_0.7.2 GenomicRanges_1.30.0 GenomeInfoDb_1.14.0 IRanges_2.12.0 S4Vectors_0.16.0 [25] BiocGenerics_0.24.0 refGenome_1.7.3 RSQLite_2.0 doBy_4.5-15 limma_3.34.3 dtplyr_0.0.2 [31] genefilter_1.60.0 RColorBrewer_1.1-2 RSkittleBrewer_1.1 plyr_1.8.4 loaded via a namespace (and not attached): [1] nlme_3.1-131 bitops_1.0-6 matrixStats_0.52.2 lubridate_1.7.1 bit64_0.9-7 [6] httr_1.3.1 tools_3.4.0 R6_2.2.2 DBI_0.7 lazyeval_0.2.1 [11] colorspace_1.3-2 mnormt_1.5-5 bit_1.1-12 compiler_3.4.0 cli_1.0.0 [16] rvest_0.3.2 xml2_1.1.1 DelayedArray_0.4.1 scales_0.5.0 psych_1.7.8 [21] digest_0.6.12 Rsamtools_1.30.0 foreign_0.8-69 XVector_0.18.0 pkgconfig_2.0.1 [26] rlang_0.1.4 readxl_1.0.0 rstudioapi_0.7 bindr_0.1 jsonlite_1.5 [31] BiocParallel_1.12.0 RCurl_1.95-4.8 GenomeInfoDbData_0.99.1 Matrix_1.2-12 Rcpp_0.12.14 [36] munsell_0.4.3 stringi_1.1.6 MASS_7.3-47 SummarizedExperiment_1.8.0 zlibbioc_1.24.0 [41] grid_3.4.0 blob_1.1.0 crayon_1.3.4 lattice_0.20-35 Biostrings_2.46.0 [46] haven_1.1.0 splines_3.4.0 hms_0.4.0 biomaRt_2.34.0 glue_1.2.0 [51] modelr_0.1.1 cellranger_1.1.0 gtable_0.2.0 assertthat_0.2.0 xtable_1.8-2 [56] survival_2.41-3 GenomicAlignments_1.14.1 memoise_1.1.0 bindrcpp_0.2 --

[lawremi]

I'm unable to reproduce this. It may be a regional thing, where different UCSC mirrors or localizations return slightly different HTML.

You could do:

debug(rtracklayer:::ucscTracks, sig="UCSCSession")

And then evaluate the first line, to assign the tracks object. Then print tracks and put the output in a pastebin or something.

[marencc]

Hi! if I do that, this is the output I get:

function (object, form = list()) { tracks <- ucscGet(object, "tracks", form) nodes <- getNodeSet(tracks, "//select/option[@selected]/text()") trackModes <- sapply(nodes, xmlValue) nodes <- getNodeSet(tracks, "//select/@name") trackIds <- unlist(nodes) nodes <- getNodeSet(tracks, "//select/../a/text()") nms <- sapply(nodes, xmlValue) names(trackIds) <- sub("^ ", "", nms[nms != "new"]) new("ucscTracks", ids = trackIds, modes = trackModes)}

[lawremi]

Did you read the part of the instructions after the debug() call? You'll need to evaluate the first line and print the tracks object.

[marencc]

I don't get any output after the debug() call. I'm sorry, I don't understand what you mean by printing the tracks object. How can I do this to help with the improvement?

[marencc]

`

debug(rtracklayer:::ucscTracks, sig="UCSCSession") refSeq <- makeTxDbFromUCSC(genome="hg19",tablename="knownGene") Tracing function ".local" in package "rtracklayer" Tracing .local(object, ...) step 2 Called from: eval(expr, p) Browse[1]> n debug: tracks <- ucscGet(object, "tracks", form) Browse[2]> Q`

[lawremi]

Instead of typing Q just type tracks.

[marencc]

Here: `

debug(rtracklayer:::ucscTracks, sig="UCSCSession") refSeq <- makeTxDbFromUCSC(genome="hg19",tablename="knownGene") Tracing function ".local" in package "rtracklayer" Tracing .local(object, ...) step 2
Called from: eval(expr, p) Browse[1]> n debug: tracks <- ucscGet(object, "tracks", form) Browse[2]> tracks Error: object 'tracks' not found `

[lawremi]

Sorry, you will need to type n twice, not just once, before that.

[marencc]

It's strange, I can't. It's automatically done.

`

debug(rtracklayer:::ucscTracks, sig="UCSCSession") refSeq <- makeTxDbFromUCSC(genome="hg19",tablename="knownGene") Tracing function ".local" in package "rtracklayer" Tracing .local(object, ...) step 2
Called from: eval(expr, p) Browse[1]> n debug: tracks <- ucscGet(object, "tracks", form)

Browse[2]> | Browse[2]> Browse[2]>

`

[marencc]

I f I try to eval (tracks) it can't find the object. if I call the "makeTxDbFromUCSC". I get the output that I sent.

[marencc]

I've restarted R and... and loaded new libraries (TxDb.Hsapiens.UCSC.hg19.knownGene) and now is working. Many thanks for your help.

[nbahti]

@marencc can you please list the steps you did to get makeTxDbFromUCSC work?

[swvanderlaan]

Yes! Please @marencc do list the steps you took to get makeTxDbFromUCSC to work.

I did:

source("https://bioconductor.org/biocLite.R")
biocLite("TxDb.Hsapiens.UCSC.hg19.knownGene")
biocLite("GenomicFeatures")
biocLite("rtracklayer")

# What is available?
supportedUCSCtables(genome = "hg19", url = "http://genome.ucsc.edu/cgi-bin/")

But I still get the same error as you @marencc. Admittedly, I also get a message stating R Studio can't install "RMySQL"... and thus I end up in dependency hell. Would love to get to know a method that doesn't end up into that hell... :-)

Thanks!

[marencc]

@swvanderlaan, @nbahti I'm sorry, I made it work for one week or so. Currently, I get this output: > refSeq2 <- makeTxDbFromUCSC(genome="hg19",tablename="knownGene") Download the knownGene table ... Error: Bad Request So, it's not working again. Previously, I just updated the TxDb.Hsapiens.UCSC.hg19.knownGene_3.2.2 I can't make it work now...

[lawremi]

Possibly just a temporary networking issue? I can't reproduce this. It's also possible that it is mirror-dependent. UCSC sometimes behaves slightly differently depending on the mirror, and thus your region of the world.

@swvanderlaan are you using R release or devel? Failure to install RMySQL probably means that there is no CRAN binary available for your version of R.

[maximilianh]

Our mirrors are identical Mysql servers. As far as I know, we haven't changed the refGene table, I think we have never changed since they were released, to avoid problems like this. Let me know if you run into similar problems again.

[lawremi]

Is anyone still experiencing issues like these?

[nbahti]

Yes, this problem showed up again since last week for me. Two or three months ago, I fixed it, but can't remember now how I did it.

On 7 May 2018 at 12:24, Michael Lawrence notifications@github.com wrote:

Is anyone still experiencing issues like these?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/lawremi/rtracklayer/issues/4#issuecomment-387139194, or mute the thread https://github.com/notifications/unsubscribe-auth/AV4oTzvoERn19VzNuPoYTVCMZZSc6EpCks5twINngaJpZM4Q9QPO .

[maximilianh]

If you're parsing our html, maybe whoever wrote the code for that can contact us one day or start a new ticket for a discussion. To import data, there shouldn't be any need for parsing html pages.

[lawremi]

If it showed up last week then it is probably the same as #8, which is fixed. So I think I'll close this issue.

Yes, parsing the HTML is pretty dumb when getting data, and I take all the blame for it. Again, the package began life as a way to programmatically manipulate the browser view. Now it's just used for the once side-feature of getting data.

[maximilianh]

Well, if you ever want this to use the mysql server or some other way (e.g. we could write a little API on our side), let me know. I'm pretty sure we can find you a nice SQL query to make it more stable and avoid you having to patch this the time.

We certainly see many http requests come in with the rtracklayer user agent. We slow them down if there are too many from the same IP address, but until now it hasn't caused too much trouble.

[lalchungnunga]

Hi, The following error occur when trying to download hg19 known gene. I am using rtracklayer 1.38.3

hg19db <- makeTxDbFromUCSC(genome = "hg19",tablename ="knownGene") Error in names(trackIds) <- sub("^ ", "", nms[nms != "new"]) : 'names' attribute [212] must be the same length as the vector [211]

[lawremi]

Please update to latest Bioc release.