Search
Question: Error through importing (project = 'SRP009615') or even to download any BigWig File
0
16 months ago by
France/Marseille
Mustafa ABUELQUMSAN0 wrote:

Hi Leonardo,

I thank you about your guidances. But i still have problematic through import BigWig file or any Project as  project = 'SRP009615'.

Despite that i check the version of "rtracelayer" that is (1.34.2 for bioc-release), furthermore i have been tried to apply all commands which is (bioc-devel mailing list   "https://stat.ethz.ch/pipermail/bioc-devel/2016-May/009336.html, ")

like the following

 1- x <- import.bw('http://duffel.rail.bio/recount/DRP000366/bw/DRR000897.bw',

+ as = 'RleList'

2- url <- unshorten_url('http://duffel.rail.bio/recount/DRP000366/bw/DRR000897.bw')

4-bw <- BigWigFile(url)

5- x <- import.bw(url, as = 'RleList') ; x <- import.bw(bw, as = 'RleList')

6- download.file('http://duffel.rail.bio/recount/DRP000366/bw/DRR000897.bw,

+ destfile = 'DRR000897.bw', method = 'curl', extra = '-L')

HERE it told me i don't have 'curl'

that processed the same my case but until this moment i couldn't to fix that problem which said me

[[[2017-02-28 16:59:34 loadCoverage: loading BigWig file http://duffel.rail.bio/recount/SRP009615/bw/mean_SRP009615.bw
Too many pushAbortHandlers, can only handle 11
Show Traceback
Error in seqinfo(con) : UCSC library operation failed]]]

i will be very thankful to guide me what is the optimal solution concerning this problem.

thanks

Mustafa.

ADD COMMENTlink
modified 16 months ago by Leonardo Collado Torres600 • written 16 months ago by Mustafa ABUELQUMSAN0

Hi Mustafa,

Can you add the full session information output please?

Thanks,

Leonardo

ADD REPLYlink written 16 months ago by Leonardo Collado Torres600

Hi Mustafa,

I can't reproduce your problem with recount 1.0.12 and rtracklayer 1.34.2. The thing that comes to mind from the bioc-devel mailing list thread you linked to is the openssl. In any case, please post the full session information.

## Un-evaluated code

library('rtracklayer')
library('devtools')
library('recount')

## Try importing a bigwig file with rtracklayer
x <- import.bw('http://duffel.rail.bio/recount/DRP000366/bw/DRR000897.bw',as = 'RleList')
x

## From http://bioconductor.org/packages/release/bioc/vignettes/recount/inst/doc/recount-quickstart.html#define-expressed-regions
## Define expressed regions for study SRP009615, only for chromosome Y
regions <- expressed_regions('SRP009615', 'chrY', cutoff = 5L,
maxClusterGap = 3000L)
regions

## Reproducibility information
Sys.time()
proc.time()
options(width = 120)
session_info()

## Evaluated code

Best,

Leonardo

ADD REPLYlink modified 16 months ago • written 16 months ago by Leonardo Collado Torres600

Hi Mustafa,

I saw that you posted the session information as a new question at my session information regarding " Error through importing (project = 'SRP009615') or even to download any BigWig File". In the future, please use the "add reply" button on the original thread. Otherwise it makes it hard for anyone else in the future to follow the set question, replies and comments to understand the solution to the problem if they encounter the same error you reported.

I know that the character limit is a bit annoying, but you can add long output like the session information in a gist https://gist.github.com/ and then post the gist link in the reply (using "text" formatting). If you do so, the website will automatically show the contents of your gist, just like it did with mine in my previous reply.

As for the actual error, your first screenshot shows a different error (or part of it, I can't see it all). I think that the error is related to the openssl system dependency in rtracklayer. I'll ask about it in bioc-devel and will get back to you.

Best,

Leonardo

ADD REPLYlink modified 16 months ago • written 16 months ago by Leonardo Collado Torres600

Mustafa,

Please add in a reply your openssl and pkg-util versions by running in your terminal (not the R one, but the unix one):

openssl version
pkg-config --version

Here is the output in two machines I have access to:

$openssl version OpenSSL 0.9.8zh 14 Jan 2016$ pkg-config --version
0.29.1

$openssl version OpenSSL 1.0.1e-fips 11 Feb 2013$ pkg-config --version
0.23

Best,

Leonardo

ADD REPLYlink written 16 months ago by Leonardo Collado Torres600

Hi Leonardo,

This is my configuration for openssl.

root@Dell-CG1XF82:~# openssl version OpenSSL 1.0.2g  1 Mar 2016 root@Dell-CG1XF82:~# pkg-config --version 0.29.1 root@Dell-CG1XF82:~# 

Best,

Mustafa.

ADD REPLYlink modified 16 months ago • written 16 months ago by Mustafa ABUELQUMSAN0

<script src="&lt;a href=" elqumsan="" a03e5f207774abca772b6bec2138a604"="">elqumsan/a03e5f207774abca772b6bec2138a604"></script>

ADD REPLYlink written 16 months ago by Mustafa ABUELQUMSAN0

Hi Leonardo,

unlimited thanks for yours instruction to put my session information here, Because of it have a bit annoying limit.

Best.

Mustafa.

ADD REPLYlink written 16 months ago by Mustafa ABUELQUMSAN0

Hi Mustafa,

If you go to your browser and enter http://duffel.rail.bio/recount/DRP000366/bw/DRR000897.bw, can you download the bigwig file to your computer? Maybe the server with the bigwig files is blocking your access. It's just a thought.

Best,

Leonardo

ADD REPLYlink written 16 months ago by Leonardo Collado Torres600

Also, what happens if you type in R the following code?

library('RCurl')
curlVersion()
ADD REPLYlink written 16 months ago by Leonardo Collado Torres600

Hi Leonardo,

I really appreciated your efforts in order to i finally arrived to the good solution. but when i run your last code which is

Thanks a lot Leonardo.

Best.

Mustafa

ADD REPLYlink written 16 months ago by Mustafa ABUELQUMSAN0

Hi Leonardo,

I tried  to download that specified file which is you tell me to download it from the link

http://duffel.rail.bio/recount/DRP000366/bw/DRR000897.bw

But also when i attempt read that file locally from my computer i faced furthermore it another error message which is

getwd()
list.files()
bigwig <- file.path("/home/abuelqumsan/Dropbox/Mustafa_Abu_Al-Qumsan_demande_Dispense_master_2015-06/scientific/discussion_notes", "DRR000897.bw" )
regions <- expressed_regions('bigwig', 'chrY', cutoff = 5L,
maxClusterGap = 3000L)

i getting this another error message

Error: nrow(url_table) > 0 is not TRUE

now i frustrating and i don't know what and how i would continuing with "recount' whereas i need to continue to conduct my project which is new and followed stage to your great work .

Thank you so much Leonardo.

Best.

Mustafa.

ADD REPLYlink written 16 months ago by Mustafa ABUELQUMSAN0

Yeah, that's not going to work. You are passing a path to a bigwig file to the "project" argument in expressed_regions(). Please check the help page of expressed_regions() and in particular check the "outdir" argument. You already had an issue of being frustrated by not following instructions and having a mix of Bioc 3.3 and 3.4 (current release) packages, which we discussed via email. Your last comment is basically a new question, so in the future please post it as one. Imagine that a new user is running into the specific problem from your last comment. If they started reading this thread they wouldn't see how it's related to their problem, unless they read everything.

Ok, so your IP is not blocked since you can download the bigwig files. recount is coded so that it can use files already downloaded following a specific file structure. I recommend using the function download_study(), which will create the particular file structure needed. Otherwise, you have to manually mimic the portion DRP000366/bw/DRR000897.bw of http://duffel.rail.bio/recount/DRP000366/bw/DRR000897.bw.

library('recount')
## Note the default of the "outdir" argument in download_study() is the "project"
download_study('DRP000366', type = 'phenotype')
download_study('DRP000366', type = 'mean')
## Check that the files are there
dir('DRP000366', include.dirs = TRUE, recursive = TRUE)
## Use the outdir argument in expressed_regions() to use the local data
regions <- expressed_regions('DRP000366', 'chrY', cutoff = 5L,
maxClusterGap = 3000L, outdir = 'DRP000366')
options(width = 120)
devtools::session_info()
ADD REPLYlink modified 16 months ago • written 16 months ago by Leonardo Collado Torres600

Hi Mustafa,

recount versions 1.0.15 and 1.1.22 now provide a more informative error in case that the 'project' argument is invalid. Details at https://github.com/leekgroup/recount/commit/ab9ca78a7ddc44247f5fbc7101b6dbdc039485ce and https://github.com/Bioconductor-mirror/recount/commit/9227ec760258eb61f57a91d7c099b7c15c942247.

Best,

Leonardo

ADD REPLYlink written 16 months ago by Leonardo Collado Torres600
0
16 months ago by
United States
Leonardo Collado Torres600 wrote:

Hi Mustafa,

This is not going to be a satisfying answer, but at least we know by now that workaround exists for your situation.

Basically, the issue you are reporting is not a recount or derfinder issue, but an rtracklayer one. For some reason,

library('rtracklayer')
x <- import.bw('http://duffel.rail.bio/recount/DRP000366/bw/DRR000897.bw',as = 'RleList')

does not work for you in Bioc 3.4 with the latest packages installed as well as recent pkg-config, openssl and a working RCurl. recount::expressed_regions() tries by default to access the BigWig files from the web and does so by using some functions in derfinder, which in turn rely on rtracklayer::import.bw(). However, recount::expressed_regions() can access the data on a local disk which removes the web issues. This is not the default behavior because it involves downloading the BigWig files even if you are interested in a small portion of the genome. The following should then work for you:

library('recount')
regions <- expressed_regions('DRP000366', 'chrY', cutoff = 5L,
maxClusterGap = 3000L, outdir = 'DRP000366')

instead of the default

regions <- expressed_regions('DRP000366', 'chrY', cutoff = 5L,
maxClusterGap = 3000L, outdir = NULL)

When specifying outdir = 'DRP000366', recount::expressed_regions()will check if the data exists and if it's missing (which would be the case normally), then it will use recount::download_study() to download the necessary files. Similarly, use outdir = 'DRP000366' when using recount::coverage_matrix().

Another option also exists, which is using SciServer Compute to access the files via a R Jupyter Notebook, run the computations on SciServer, save your files (I assume that you plan to run recount::coverage_matrix() next), then download them to your computer. Most of this is described in the recount vignette at http://bioconductor.org/packages/release/bioc/vignettes/recount/inst/doc/recount-quickstart.html#accessing-recount-via-sciserver.

If you are interested in accessing the BigWig files from the web, please post a new question using the rtracklayer tag. Basically, with the output of:

library('rtracklayer')
x <- import.bw('http://duffel.rail.bio/recount/DRP000366/bw/DRR000897.bw',as = 'RleList')
options(width = 120)
devtools::session_info()

I have reported most of this already at the bioc-devel mailing list https://stat.ethz.ch/pipermail/bioc-devel/2017-March/010518.html and have run out of ideas on what could be the source of the problem behind your rtracklayer issue.

Best,
Leonardo

ADD COMMENTlink written 16 months ago by Leonardo Collado Torres600
Please log in to add an answer.

Content
Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 256 users visited in the last hour