ToDo

• add useful links
• add links to important papers
• add info on how to connect to the cluster and jupyterlab
• directory structure
• expand useful file locations
• create tutorials (on a separate repository?)

Useful links

• CAST website
• UKBB user site
• AllOfUs researcher workbench
• AllOfUs data browser
• Frazer lab cluster wiki

Reading

Coming soon.

Getting started

Initially you should have a user account on the Frazer Lab cluster. In order to access the cluster from your home terminal:

ssh [email protected]

You should typically not work on the head node. After you have logged in, to navigate elsewhere:

qlogin

To request more memory/cores:

qlogin -pe smp 4 -l h_vmem=4G # request 4 cores (4 Gb/core)

Connect to jupyterlab:

jupyterlab

To delete jupyterlab jobs (via command line):

1. Run qstat to find the job ID

qstat

{
job-ID  prior   name       user         state submit/start at     queue                          slots ja-task-ID
-----------------------------------------------------------------------------------------------------------------
8803323 0.55617 jupyterhub username1    r     06/06/2022 11:37:43 juphigh.q@fl-n-1-10               16
8803325 0.55617 jupyterhub username2    r     06/06/2022 14:08:13 juphigh.q@fl-n-1-3                16
}

2. qdel your job ID

qdel 8803323

Install software

Add software to ~./bashrc:

export PATH=/frazer01/software/bcftools-1.9/bin:$PATH
export PATH=/frazer01/software/bedtools-2.27.1/bin:$PATH
export PATH=/frazer01/software/samtools-1.9/bin:$PATH
export PATH=/frazer0l/software/tabix-0.2.6/bin:$PATH

miniconda

From these instructions. First you will need to install miniconda locally. If this is something you've never done before or have limited experience with, it might be useful to ask someone for more help with this issue. Otherwise proceed as follows.

Download the most recent version of miniconda (assuming you are installing this in your home directory):

cd ~
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh

Follow the prompts on the screen to finish the installation.

To make the changes take effect, close and then re-open your terminal window.

Test your installation. In your terminal window or Anaconda Prompt, run the command conda list. A list of installed packages appears if it has been installed correctly.

Plink

For very obnoxious reasons both the original plink (1.9) and plink 2.0 are probably required. Installing them locally is an option, but it may just make an alias. Do this by adding the following two lines to your ~./bashrc file:

alias plink2="/frazer01/software/plink-2.3/plink2_64"
alias plink="/frazer01/software/plink-1.90b3x/plink"

Set up a python kernel for the jupyter lab

Coming soon.

Running a qsub job

Run qsub:

qsub my_script.sh

You can include the following to the SH file to pass to qsub as:

#!/bin/bash

#$ -N job_name
#$ -pe smp 4                                ### specify number of cores requested
#$ -l h_vmem=4G                             ### specify amount of memory per core (default is 4Gb per core)
#$ -l short                                 ### specify queue ([short, week, long, opt], default is all)
#$ -V                                       ### export your current environment parameters to the job
#$ -cwd                                     ### change the working directory to where the script was submitted from
#$ -e ~/std.err                             ### redirect stderr to this file
#$ -o ~/std.out                             ### redirect stdout to this file
#$ -t 1-10                                  ### define array jobs (in this case will run 10 jobs), use $SGE_TASK_ID to get access to the array index
#$ -l hostname="fl-n-1-1|fl-n-1-3|fl-n-1-4" ### define the nodes you want your jobs to run (1-17)

Always make sure that environment is activated in the SH file:

source /home/username/.bashrc

Useful File Locations

• UKBB pgen files: /frazer01/projects/CEGS/analysis/apoe_haplotypes/input/genotypes/ukbb
• UKBB subjects by ethnicity: /frazer01/projects/CEGS/analysis/ukbb_hla_type_gwas/pipeline/ethnicity/subjects_by_ethnicity/

Forking a Github repo branch

First, go on GitHub and click fork in the upper right hand side. This should add the repository to your user account homepage.

git clone <URL-of-repository>

Navigate into directory on your local machine or server.

git status

Should say "on branch master"

git branch

Should say "* master"

git pull upstream master

Should give error ‘upstream’ does not appear to be a git repo

git remote add upstream <URL>
git pull upstream master

should execute with no errors

git checkout -b desired-branch-name
git branch

Should say "* desired-branch-name". We are trying to see if this is executing the correct sequence of commands.

git add file1
git commit -m ‘Update file1’
git push origin desired-branch-name

Then go to the github website and go to PULL REQUESTS in the upper left corner. Click issue pull request and add any comments. Make sure you don’t click merge on the website until someone reviews your code!