ORCA setup

[lchong]

@tmozgach @sjackman

Hi all, we can use this issue to centralize the conversation about ORCA software installations prior to the event. Here are the requests from team leaders so far:

Project 2

See issue for full discussion (hackseq/hackseq_projects_2017#2)

• methpipe

Project 6

• Bedtools
• BWA
• FastQC
• Graphviz
• Long Ranger license prevents redistribution
• LRSIM
• makefile2graph
• Miller
• R
• RStudio Server
• samtools
• seqtk

Project 8

• scikit-learn
• pytorch

R Packages

• Bioconductor

• data.table

• devtools

• DiagrammeR

• doMC

• DT

• fuzzyjoin

• ggplot

• igraph

• knitr

• pander

• revealjs

• rmarkdown

• roxygen2

• tidyverse

• phytools

I'll follow up shortly with the leaders to remind them to put together their lists.

[sjackman]

@tmozgach Please brew install graphviz lrsim makefile2graph
rstudio-server has two open issues: https://github.com/Homebrew/homebrew-science/pull/5502 and https://github.com/Homebrew/homebrew-science/pull/6315

[BrunoGrandePhD]

What packages are installed for R? Since installing several R packages on Linux can take a while due to compilation, it would help if a few main ones come pre-installed. I've put together an initial list below, but it might get longer after discussing more specific goals with team members.

R Packages

• tidyverse
• rmarkdown
• DT
• fuzzyjoin
• pander
• doMC
• revealjs
• devtools
• roxygen2
• data.table

[sjackman]

There are currently no R packages installed, other than the base packages.
@tmozgach Can you please add these R packages to the ORCA image, and also ggplot2 and knitr? You can install them with

Rscript -e 'install.packages(c("tidyverse", …, "data.table", "ggplot2", "knitr"))'

and Bioconductor, using these instructions: https://www.bioconductor.org/install/

Rscript -e 'source("https://bioconductor.org/biocLite.R"); biocLite()'

Thanks!

[morien]

Will add to the list of R packages to be installed: 'phytools', plus dependencies 'ape' (>= 4.0), and 'maps'

Git, of all things, isn't on the server software list, but I have to assume it's there... (?)

[sjackman]

@morien Only those software installed by either brew or pip is listed. git is installed by apt-get and so isn't included in this list.

[sjackman]

@tmozgach Please add igraph and DiagrammeR to the list of R packages.

[morien]

@sjackman ah okay, well then we can strike "tree" from the list of requests, that would also get installed with the OS package manager, not brew or pip

[sjackman]

The Docker image that we're using is bcgsc/orca:2017.08.23 which is based on linuxbrew/linuxbrew:1.2.8 which is based on ubuntu:trusty. The next release of ORCA will be based on linuxbrew/linuxbrew:1.3.* and ubuntu:xenial. Feel free to check it out if you'd like to poke around and see what software is installed, and try out your pipeline prior to Hackseq. Note that the image is 10 GB.

[lchong]

From Project 8: "Could we get scikit-learn and pytorch preinstalled on our ORCA machines?"

• scikit-learn
• pytorch

[sjackman]

@tmozgach I've added the software listed in the comments above the to the summary of this issue:
#33 (comment)
Can you please tick off the boxes that you've installed in the ORCA image?

[sjackman]

@lchong Could you please add Tanya @tmozgach as a collaborator on this repository, so that she can tick off the boxes in the summary of this issue?

[sjackman]

@lchong Could you please send me the names and e-mail addresses of the project leads so that we can create ORCA accounts for them?

[lchong]

@sjackman try this link:

https://docs.google.com/spreadsheets/d/1IFDkStUlfCYGfYhpHPN-TYj_OxQ2LfTJ0tMC1gQZH6Y/edit?usp=sharing

[sjackman]

Thanks, Lauren!