search

thelovelab / tximeta

Transcript quantification import with automatic metadata detection
https://thelovelab.github.io/tximeta/
66 stars 11 forks source link

Unsuccessful connection to AnnotationHub blocks tximeta on HPC #67

Closed danilotat closed 2 years ago

danilotat commented 2 years ago

Hi Mike, thanks for your work on this package. I've used it many times with success, till now when I'm facing a problem with its usage in a HPC server. Obviously I've checked for internet connection and, while checking for suggested solution outside I came across this issue on AnnotationHub. It seems to reflect my scenario, where on our cluster nslookup requests are blocked by default. Is there any existing workaround ?
I'm using tximeta=1.14.0 installed from Bioconda. Here's the stdout

NOTE: linkedTxome with source='Ensembl', ensembldb will be used to parse GTF.
this may produce errors if the GTF is not from Ensembl, or has been modified.
set useHub=FALSE in tximeta to avoid download of reference txome from AnnotationHub.
alternatively use a different string for source argument
writing linkedTxome to /hpc/../salmon_index_GRCh38_ensembl.json
saving linkedTxome in bfc (first time)
importing quantifications
reading in files with read.delim (install 'readr' package for speed up)
1 2
found matching linked transcriptome:
[ Ensembl - Homo sapiens - release 105 ]
useHub=TRUE: checking for EnsDb via 'AnnotationHub'
Cannot connect to AnnotationHub server, using 'localHub=TRUE' instead
Error in .updateHubDB(hub_bfc, .class, url, proxy, localHub) :
  Invalid Cache: sqlite file
  Hub has not been added to cache
  Run again with 'localHub=FALSE'
Calls: tximeta ... AnnotationHub -> .Hub -> .create_cache -> .updateHubDB
Execution halted
mikelove commented 2 years ago

useHub=FALSE should fix?

danilotat commented 2 years ago

It works, although I'm still seeing many errors coming from unsuccessful connections with ftp.ensemblgenomes.org: sorry for such an annoying issue. Which is the impact of using/not using AnnotationHub?

mikelove commented 2 years ago

Tximeta extends tximport in two ways: 1) it packages the matrices into a Bioconductor (S4) object and 2) it automatically downloads metadata based on inferring the provenance of the transcripts.

If you don’t have ftp access then (2) won’t work. You can still take advantage of (1) by setting skipMeta=TRUE.

mikelove commented 2 years ago

The AHub aspect isn’t really relevant. Tximeta works with or without AHub…

danilotat commented 2 years ago

Fine, thanks for the quick response. I’ll close this issue