Question

tximeta - problem importing salmon files

0

Entering edit mode

rihards.saksis • 0

@rihardssaksis-22942

Last seen 4.3 years ago

Hello.

After quantifying reads with "Salmon", i try to import them with "tximeta" package for use with "DESeq2", but i receive this error: "Error in download.file(url, destfile, quiet = TRUE) : cannot open URL 'http://hgdownload.cse.ucsc.edu/goldenPath/hg38/database/chromInfo.txt.gz'

And this warning message: In addition: Warning message: In .getcdsIDX(mcols0$type, mcols0$phase) : The "phase" metadata column contains non-NA values for features of type stop_codon. This information was ignored."

Salmon version: 1.2.1 Tximeta version: v1.6.2 Newest R and RStudio versions.

Normally i would ignore the warning, but the error does not permit me to continue. I believe, that the website is not working. Has it moved to a different domain or just temporary unavailable?

Thank you in advance.

tximeta deseq2 • 2.2k views

ADD COMMENT • link updated 4.4 years ago by Michael Love 43k • written 4.4 years ago by rihards.saksis • 0

score 3 · Answer 1 · 2020-07-07

3

Entering edit mode

Michael Love 43k

@mikelove

Last seen 2 days ago

United States

Thanks for the report.

tximeta uses Seqinfo from the GenomeInfoDb package to fetch the correct chromosome names and lengths from UCSC (for GENCODE transcriptomes).

It appears the link is down right now or has moved, the following also hangs for me:

> Seqinfo(genome="hg38")
Error in download.file(url, destfile, quiet = TRUE) :
  cannot open URL 'http://hgdownload.cse.ucsc.edu/goldenPath/hg38/database/chromInfo.txt.gz'
Calls: Seqinfo ... <Anonymous> -> fetch_table_from_UCSC -> fetch_table_from_url

This hasn't happened before, but it affects core packages upstream, so either it will be resolved on the UCSC side (maybe it's just down right now) or on the Bioc side (maybe they moved the location and we need to update).

In the meantime, you can use skipMeta=TRUE to import the transcript-level counts. Maybe I can work on a work-around in the case that this is the only metadata task that is broken (but e.g. the GTF file is reachable, parsed, cached locally, etc.)

ADD COMMENT • link 4.4 years ago Michael Love 43k

0

Entering edit mode

Thank you! Could it be that the function which summarizes transcriptome counts to gene level doesn`t work because of the missing file?

ADD REPLY • link 4.4 years ago rihards.saksis • 0

0

Entering edit mode

It probably won't but I'm writing a fix now, to avoid this one step that requires UCSC server connect.

ADD REPLY • link 4.4 years ago Michael Love 43k

1

Entering edit mode

I forgot, I already pushed a fixed for this last year. You can use skipSeqinfo=TRUE. See description here:

https://rdrr.io/bioc/tximeta/man/tximeta.html

ADD REPLY • link 4.4 years ago Michael Love 43k

0

Entering edit mode

Will this in any way affect the quality going forward? Should i redo it when the server comes online?

ADD REPLY • link 4.4 years ago rihards.saksis • 0

1

Entering edit mode

Your chromosomes won’t have lengths as according to UCSC. I’d redo it for completeness.

ADD REPLY • link 4.4 years ago Michael Love 43k

score 0 · Answer 2 · 2020-07-07

The download server seems to be down. But do note that tximeta is simply downloading a freely available resource, which is different from being responsible for that resource, or for that matter knowing anything about what the suppliers of that resource might be up to. That sort of question should be directed to the folks at UCSC Genome Browser.