search

pachterlab / sleuth

Differential analysis of RNA-Seq
http://pachterlab.github.io/sleuth
GNU General Public License v3.0
305 stars 94 forks source link

Sleuth_live splitting Ensemble gene ID's into columns... #234

Open zxc098 opened 4 years ago

zxc098 commented 4 years ago

Hello,

The actual Sleuth package and plots within sleuth_live work/look great, but for some reason it appears that in the tables (when attempting to use sleuth_live), every single character from the Ensemble gene ID is being read into every single column, one character per column (screen shot below). Screen Shot 2019-11-12 at 12 23 21 PM

If I make gene tables or use plots in Sleuth itself (i.e., just use the various functions in R), everything works great, and moreover, if I download the table from sleuth_live, it looks correct (i.e., the ensemble ID is not split over the columns)...but for some reason it's reading into/appearing in sleuth_live in this weird way.

I have tried coercing the data/columns every which way in Sleuth, as well as saving sleuth object in different formats and reading back in to launch sleuth_live, but nothing seems to work...

I am using Mac OS, with latest versions of R and R studio.

I may be barking up the wrong tree here, since this could be some kind of weird Shiny issue, but I just don't know enough/have enough experience with Shiny to figure it out...any help would be greatly appreciated,

Thank you.

vaofford commented 4 years ago

Hi,

I'm having a similar issue with R v3.6, sleuth v0.30.0 and shiny v0.14.

Screen Shot 2019-11-18 at 10 24 10

This is using a script and data which hasn't changed, just a fresh install of the R packages.

Many thanks,

Victoria

sigusn commented 4 years ago

I´m having the same issue, old scripts, just fresh analysis and the tables are all messed up.

vaofford commented 4 years ago

We got around this using:

Roll back htmlwidgets

wget https://www.stats.bris.ac.uk/R/src/contrib/Archive/htmlwidgets/htmlwidgets_1.3.tar.gz Rscript -e "install.packages('htmlwidgets_1.3.tar.gz', repos = NULL, type='source')"

Roll back htmltools

wget https://www.stats.bris.ac.uk/R/src/contrib/Archive/htmltools/htmltools_0.3.tar.gz Rscript -e "install.packages('htmltools_0.3.tar.gz', repos = NULL, type='source')"

Get the development version of shiny

Rscript -e "devtools::install_github('rstudio/shiny')"

Seems like it's linked to: https://github.com/rstudio/htmltools/issues/143

jlewandowski commented 4 years ago

I am having the same issue too. @vaofford I am not totally sure I understand the fix. Would it be possible to list out the steps needed / have others gotten around this?

pdru commented 4 years ago

Hello, same problem too, the fix (downgrade) worked sometimes (not often), not systematically at all, and not reproducible (tried on 2 mac, one ok for one session, same code (with downgrade) didn't worked. Tested with 2 different sets of data, on OSX, PC...(on of the set worked well on november...

sophie-harrington commented 4 years ago

Hi, also having the same problem here. I haven't got the downgrading to work unfortunately. Any other ideas? Thanks!