TCGAbiolinksGUI has error when running DEA analysis

[AmyOlex]

Hello! I just found your tool that looks amazing, but I'm having some problems getting it to run a DEA analysis. I have R 3.3.3 and I'm unable to install TCGAbiolinksGUI via Bioconductor, so I followed the directions from here where on the last step I used the current package on bioconductor (TCGAbiolinksGUI_1.0.0.tar.gz): https://libraries.io/github/BioinformaticsFMRP/TCGAbiolinksGUI

It seems to be working and I can download and prep data, but when I try to run a DEA analysis it throws the following error:

Warning: Unknown or uninitialised column: 'disease_type'. Warning: Unknown or uninitialised column: 'disease_type'. Warning: Unknown or uninitialised column: 'disease_type'. Warning: Unknown or uninitialised column: 'disease_type'. Warning: Unknown or uninitialised column: 'disease_type'. Warning: Error in cbind: object 'dataFilt' not found ERROR: [on_request_read] connection reset by peer

I tried the Docker image, but it has the same problem and disconnects from the server when running a DEA analysis.

My R session info is the following:

sessionInfo() R version 3.3.3 (2017-03-06) Platform: x86_64-apple-darwin13.4.0 (64-bit) Running under: OS X Mavericks 10.9.5

locale: [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages: [1] grid parallel stats4 stats graphics grDevices utils datasets methods
[10] base

other attached packages: [1] dplyr_0.5.0
[2] data.table_1.10.4
[3] readr_1.1.0
[4] googleVis_0.6.2
[5] ELMER_1.4.1
[6] ELMER.data_1.4.0
[7] Homo.sapiens_1.3.1
[8] TxDb.Hsapiens.UCSC.hg19.knownGene_3.2.2
[9] GO.db_3.4.0
[10] OrganismDbi_1.16.0
[11] GenomicFeatures_1.26.4
[12] IlluminaHumanMethylation450kanno.ilmn12.hg19_0.6.0 [13] minfi_1.20.2
[14] bumphunter_1.14.0
[15] locfit_1.5-9.1
[16] iterators_1.0.8
[17] foreach_1.4.3
[18] Biostrings_2.42.1
[19] XVector_0.14.1
[20] htmlwidgets_0.8
[21] ggrepel_0.6.5
[22] stringr_1.2.0
[23] plotly_4.6.0
[24] ggplot2_2.2.1
[25] clusterProfiler_3.2.14
[26] DOSE_3.0.10
[27] reshape2_1.4.2
[28] pathview_1.14.0
[29] org.Hs.eg.db_3.4.0
[30] AnnotationDbi_1.36.2
[31] SummarizedExperiment_1.4.0
[32] Biobase_2.34.0
[33] GenomicRanges_1.26.4
[34] GenomeInfoDb_1.10.3
[35] IRanges_2.8.2
[36] S4Vectors_0.12.2
[37] BiocGenerics_0.20.0
[38] shinyjs_0.9.0.9001
[39] shinyBS_0.62
[40] TCGAbiolinks_2.3.23
[41] shinyFiles_0.6.2
[42] shiny_1.0.3
[43] TCGAbiolinksGUI_1.0.0
[44] shinydashboard_0.5.3
[45] BiocInstaller_1.24.0

loaded via a namespace (and not attached): [1] R.utils_2.5.0 RSQLite_1.1-2 trimcluster_0.1-2
[4] BiocParallel_1.8.2 devtools_1.12.0 DESeq_1.26.0
[7] munsell_0.4.3 codetools_0.2-15 preprocessCore_1.36.0
[10] miniUI_0.1.1 withr_1.0.2 colorspace_1.3-2
[13] GOSemSim_2.0.4 knitr_1.15.1 supraHex_1.12.0
[16] robustbase_0.92-7 git2r_0.18.0 KEGGgraph_1.32.0
[19] KMsurv_0.1-5 mnormt_1.5-5 hwriter_1.3.2
[22] downloader_0.4 c3net_1.1.1 ggthemes_3.4.0
[25] EDASeq_2.8.0 diptest_0.75-7 R6_2.2.0
[28] doParallel_1.0.10 illuminaio_0.16.0 RJSONIO_1.3-0
[31] flexmix_2.3-14 bitops_1.0-6 reshape_0.8.6
[34] fgsea_1.0.2 assertthat_0.2.0 scales_0.4.1
[37] nnet_7.3-12 gtable_0.2.0 affy_1.52.0
[40] genefilter_1.56.0 cmprsk_2.2-7 GlobalOptions_0.0.11
[43] splines_3.3.3 rtracklayer_1.34.2 lazyeval_0.2.0
[46] GEOquery_2.40.0 hexbin_1.27.1 selectr_0.3-1
[49] broom_0.4.2 httpuv_1.3.3 qvalue_2.6.0
[52] RBGL_1.50.0 tools_3.3.3 psych_1.7.3.21
[55] nor1mix_1.2-2 affyio_1.44.0 RColorBrewer_1.1-2
[58] siggenes_1.48.0 Rcpp_0.12.10 plyr_1.8.4
[61] zlibbioc_1.20.0 purrr_0.2.2 RCurl_1.95-4.8
[64] ggpubr_0.1.2 openssl_0.9.6 GetoptLong_0.1.6
[67] viridis_0.4.0 zoo_1.8-0 cluster_2.0.6
[70] magrittr_1.5 dnet_1.0.10 DO.db_2.9
[73] circlize_0.3.10 colourpicker_0.3 survminer_0.3.1
[76] mvtnorm_1.0-6 whisker_0.3-2 matrixStats_0.52.2
[79] aroma.light_3.4.0 mime_0.5 hms_0.3
[82] xtable_1.8-2 minet_3.32.0 XML_3.98-1.6
[85] mclust_5.2.3 gridExtra_2.2.1 shape_1.4.2
[88] biomaRt_2.30.0 tibble_1.3.0 R.oo_1.21.0
[91] htmltools_0.3.6 tidyr_0.6.1 geneplotter_1.52.0
[94] DBI_0.6-1 matlab_1.0.2 ComplexHeatmap_1.12.0
[97] MASS_7.3-47 fpc_2.1-10 ShortRead_1.32.1
[100] Matrix_1.2-8 parmigene_1.0.2 quadprog_1.5-5
[103] R.methodsS3_1.7.1 igraph_1.0.1 km.ci_0.5-2
[106] GenomicAlignments_1.10.1 registry_0.3 foreign_0.8-68
[109] xml2_1.1.1 annotate_1.52.1 rngtools_1.2.4
[112] pkgmaker_0.22 multtest_2.30.0 beanplot_1.2
[115] rvest_0.3.2 doRNG_1.6.6 digest_0.6.12
[118] ConsensusClusterPlus_1.38.0 graph_1.52.0 base64_2.0
[121] fastmatch_1.1-0 survMisc_0.5.4 dendextend_1.5.2
[124] edgeR_3.16.5 curl_2.6 kernlab_0.9-25
[127] Rsamtools_1.26.2 modeltools_0.2-21 rjson_0.2.15
[130] nlme_3.1-131 jsonlite_1.4 viridisLite_0.2.0
[133] limma_3.30.13 lattice_0.20-35 KEGGREST_1.14.1
[136] httr_1.2.1 DEoptimR_1.0-8 survival_2.41-3
[139] png_0.1-7 prabclus_2.2-6 Rgraphviz_2.18.0
[142] class_7.3-14 stringi_1.1.5 latticeExtra_0.6-28
[145] memoise_1.1.0 ape_4.1

[labrazil]

Which tumor type were you investigating?

[labrazil]

Also, please consider running the latest R (R version 3.4.0 (2017-04-21))

[torongs82]

Hi @AmyOlex thank you for your interest in the tool. I agree about knowing the tumor type so we can try to reproduce the error. Anyway It seems that 'ERROR: [on_request_read] connection reset by peer' can depends also by the browser you are using, can you try firefox or chrome?

[AmyOlex]

Thanks for the quick reply. I am studying HNSC. I was just running a small test so I downloaded all the HNSC SolidTumorNormal gene expression HTseq-count data (44 samples) from the harmonized GDC. I went to run a DEA and selected to normalize by geneLeangth, then set the analysis parameters to FC=0.5, pval = 0.05, group column = gender, group 1 = male, group 2 = female. It fails regardless of the DEA algorithm type. This same thing happens when I run the analysis using the Docker image and that is using R 3.3.2, so a bit older than my version.

I've tried Google Chrome and Firefox and they both have the same problem. I'm going to try ACC to see if that works.

Also, When I enter query parameters and hit Visualize Data I get no pie charts. I have to download the data first, then the charts show up. Is that the way it is supposed to work?

Thanks! -Amy

Amy L. Olex Bioinformatics Specialist C. Kenneth and Dianne Wright Center for Clinical and Translational Research

Virginia Commonwealth University 1200 E. Clay Street P.O. Box 980261 Richmond, VA 23298

Phone: 804-828-1621 Fax: 804-628-4723

On Fri, Apr 28, 2017 at 10:24 AM, Antonio Colaprico < notifications@github.com> wrote:

Hi @AmyOlex https://github.com/AmyOlex thank you for your interest in the tool. I agree about knowing the tumor type so we can try to reproduce the error. Anyway It seems that 'ERROR: [on_request_read] connection reset by peer' can depends also by the browser you are using, can you try firefox or chrome?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/BioinformaticsFMRP/TCGAbiolinksGUI/issues/13#issuecomment-298011947, or mute the thread https://github.com/notifications/unsubscribe-auth/AHQGLkT-1tbwPG8WJ_bx8bW6dLKW8979ks5r0faegaJpZM4NLiuf .

[tiagochst]

Thanks for the report, there is a bug.

Just to illustrate better the problem:

Data

[tiagochst]

About the graphs, some time is required to access GDC database, so the plot will only be shown if visualize data is pressed. There will be no plot if the query results in "No results found". I just realized the notifications is not working as it should with R 3.4 if it is in the middle of the code. Unfortunately, we will need to replace the shinyBS package for this notifications

[AmyOlex]

I tried not filtering and I still have the same problem. I thought it may be a browser issue like the visualization of the pie charts, but Safari, Chrome and FireFox all terminate when I try the DEA analysis. I've tried the CHOL and ACC datasets with primary tumor vs solid normal tissue. I also tried it on my Ubuntu 16 desktop using FireFox and I have the same termination problem with the DEA analysis. This time I was using the TCGABiolinksGUI instance at https://tcgabiolinksgui.shinyapps.io/tcgabiolinks/ .

[tiagochst]

That demo is an old version, could you try this one: http://tcgabiolinks.fmrp.usp.br:3838/. It has the last version of all packages.

[tiagochst]

The Safari charts problem exists, I'll open an issue on the googleVis github to see if they can help me.

[AmyOlex]

This version works! Thanks!

Is this just a demo or can I always use this site for the GUI? I am having no luck installing in on my laptop or using the Docker implementation.

Also, how much traffic can the server take? I am teaching a workshop about GDC in a while and I am considering having the group of about 30 people do an assignment with the TCGABiolinksGUI, but I don't want to jam up your server.

Thanks! -Amy

Amy L. Olex Bioinformatics Specialist C. Kenneth and Dianne Wright Center for Clinical and Translational Research

Virginia Commonwealth University 1200 E. Clay Street P.O. Box 980261 Richmond, VA 23298

Phone: 804-828-1621 Fax: 804-628-4723

On Thu, May 11, 2017 at 8:57 PM, Tiago Chedraoui Silva < notifications@github.com> wrote:

That demo is an old version, could you try this one: http://tcgabiolinks.fmrp.usp.br:3838/. It has the last version of all packages.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/BioinformaticsFMRP/TCGAbiolinksGUI/issues/13#issuecomment-300955387, or mute the thread https://github.com/notifications/unsubscribe-auth/AHQGLrrUKqs4B-XiAhdJas9YOz3oFbHWks5r465ogaJpZM4NLiuf .

[tiagochst]

It is just a demo, I would not use for analysis. Maybe it can be used for a small workshop, I'm just not sure it can handle 30 connections.

Also, I would recommend machines with more than 8Gb. Unfortunately docker is working well on mac, but for windows I had some troubles too. I hope they make it easier to use on windows.

[torongs82]

Hi @tiagochst thank you, but I was thinking that I can help you @AmyOlex with a solution. I gave a workshop with 40 people attending, and for that purpose we can offer for free of course but for a short time such as 2-3 days a link to a website with 50 different ports that are linking to different dockers with TCGAbiolinksGUI deployed on our servers. And as I made for my participants you can assign each port to each attender. If you are interested please write me a mail acolapri@ulb.ac.be and we see how we can help you. Best. Antonio.