Big Data Summer Institute 2019 | Genomics
[email protected]

• Reproductive Research Demo
• Canvas
• Slack
• BDSI Wiki
• UNIX tutorial
• Useful UNIX one-liners
• Nano tutorial
• Download MobaXterm
• Data Visualization

1. Graphical User Interface (GUI)

Enter the URL http://biostat-login.sph.umich.edu on Chrome or Firefox
Enter your password and complete Duo login
Interactive Apps >> RStudio You can also submit jobs from here.

2. Command line

Make sure you're on MWireless. ssh <yourUNIQNAME>@biostat-login.sph.umich.edu
Enter your password and complete Duo login You can use ssh -Y <yourUNIQNAME>@biostat-login.sph.umich.edu so that you can open figures interactive when you're running R from the command line. If you want to request resources for a long-running or memory intensive command, you will need to:

1. submit a job using sbatch
2. request an interactive job on a compute node using srun
   srun --time=2:00:00 --mem=2GB --pty /bin/bash

If you want to log specifically into one of the login nodes. For example, you might want to run htop in the same login node that you are running a process.

ssh <uniqname>@idran.bio.sph.umich.edu  
ssh <uniqname>@bajor.bio.sph.umich.edu  

#head to your home directory
cd ~/
#clone the repo
git clone https://github.com/bnwolford/BDSI.git
#move into the new folder that  has been  created
cd BDSI

Now you find yourself in a folder that mirrors the code and files in this repo.

Make a job array fil, for example some submit_jobs.sh, with one command per line. For example:

for f in `ls /tmp/bdsi2019/genomics/data/prs/gfg/*snps.bim`; do base=`basename $f .bim`; echo “plink --bfile /tmp/bdsi2019/genomics/data/prs/gfg/$base --score --out $base”; done > submit_score.sh

And if you cloned this repo into your home directory (see the 'Access this respository from home directory' section above) you can execute a command as follows, customizing memory, time, cpu, etc.
perl ~/BDSI/create.slurm.scripts.opts.pl -f submit_jobs.sh -m 2 -t 12:00:00 -j <name> -c 1
This script will tell you to run something like this
sbatch /home/bdsi2019/genomics/data/prs/gfg/<name>.slurm.sh
Execute that command. Check to see your jobs in the queue.
squeue -u <username>

Your questions answered here.

cd /home/bdsi2019/genomics/data/prs

To access code from the Zhou lab GitHub, you can clone the repo into your home directory.
cd ~/
git clone https://github.com/xzhoulab/DECComparison.git
cd DECComparison

We want to practice a typical workflow which involves sharing a cenral dataset and performing indpendent analyses in our own directories. For this reason, I've downloaded the data for you, and those steps are now commented out (i.e. # is placed at the front of the code line).

We want to start an interactive session to request compute resources before we get started.
srun --time=2:00:00 --mem=2GB --pty /bin/bash
In an analysis directory in your home directory, launch R with R. Follow the R code and answer the questions here.

cd /home/bdsi2019/genomics/data/popgen

cd /home/bdsi2019/genomics/data/mr