/GettingStarted

Example scripts for common GoldbergLab analyses

Primary LanguageShellMIT LicenseMIT

Getting Started

Example scripts for common analyses. We're starting this directory to keep track of example scripts we've made for new Goldberg Lab members-- please feel free to contribute anything that might be useful to future lab members.

Duke Compute Cluster quick start

DCC user guide (lots of useful info here on how to get started): https://oit-rc.pages.oit.duke.edu/rcsupportdocs/

Getting an account:

To SHH onto the DCC:

  • ssh netid@dcc-login.oit.duke.edu
  • Multifactor authentication is now required for all logins, on and off campus.
  • Your username/password is your NetID/password.

Important directories:

  • /hpc/home/netID --This is your 10GB personal home directory. There are a few shared directories where you can make your own “yourNetID/” subfolder (i.e., within each of these directories, please make yourself a "yourNetID/" directory (e.g., my (Katharine K) work directory is /work/klk37/) and work within that directory.
  • /work --Use this as a temporary location for active analyses. This is ~500TB of unpartitioned space shared across all DCC users. This space is not backed up and files in /work are automatically deleted after 75 days.
  • /hpc/group/goldberg --This is our lab directory, with 1TB of storage and 7 days of backups
  • /datacommons/goldberg --This is our lab directory for archival storage. Use this for storing important (not intermediate or temporary) files and scripts. This Datacommons directory is mounted to the cluster for easy access, but read/write won’t be as fast compared to /work or /hpc directories

IMPORTANT NOTES:

  • Raw data and scripts should always be saved to a secondary safe location, outside of the cluster.
  • A good practice is to keep all scripts in an hpc/ or datacommons/ directory, and run all analyses in work/. By keeping active analyses and large intermediate files in work/, we prevent overflowing lab storage and interrupting each others’ jobs. Just remember to always keep a copy of raw data and the scripts needed to generate downstream files and output in a backed-up location.

Summaries/walkthroughs within this directory:

  • Prep1000G_MalariaProject_summary.md written Oct 2020 to assist with looking at possible selection at malaria-interacting loci in 1000G admixed populations. contains some guidance basic PLINK commands and PCA/ADMIXTURE.
  • ADMIXTURE_allele-freq_combined.pdf written Oct 2020. Example walk-through for plotting ADMIXTURE results in R and comparing observed to expected allele frequencies.
  • Intro to Unix_Cluster Computing.pdf written Aug 2020. Notes for introduction to Unix/Terminal and the Duke Compute Cluster architecture.
  • SLiM_jobarray_example.pdf written March 2021. Notes on how to use job arrays to run and keep track of many SLiM simulations on the cluster.
  • JupyterNotebook_DCC_Container.pdf written Sept 2021. Notes on building Singularity image for software containerization & for running Jupyter Notebook on the DCC cluster from a local browser (can be extended to running Rstudio).

Public datasets:

Software: