Stat/Biostat 550

Getting Started:

This information is modified from the lab information first developed by Amy Anderson, 2001. This version is updated from 2006, but not everything regarding Biostat computing may be current.

There are are a few things you should do before you begin the Labs for Stat550, or other StatGen computing. Start by going to the class web page at http://www.stat.washington.edu/thompson/Stat550/
Scroll down the page until you see the Computing section.

  1. If you don’t already have one, you need to get a biostat account. Your biostat account gives you access to the machines that contain the programs we will be using this quarter.  To get a biostat account, you just need to email the Biostatistics computing staff (bite@u.washington.edu), and ask for one. You need to give your full name and your UWNetID (the name on your u.washington account). (Note: I will collect a list in the first class, and send it in.)
  1. Make sure you can access the StatGen computing web pages. These web pages are maintained on the central UW system and are password-protected. That is, you will need to use your UWNet ID to access them.

Once you have a biostat account, you can proceed to step three:

 

  1. Log in to the computer named abacus.biostat.washington.edu. using an SSH program. If you are starting out on a Linux computer (for example, one of the Statistics department Linux servers), you can ssh into the Biostat machines by typing:

    % ssh name@abacus.biostat.washington.edu

    where name is your user name for your biostat account (don’t actually type the % sign—it’s just there to denote the prompt that should already be sitting on your screen). Your prompt sign may be a ``$'' sign instead of a %, or something else. From a Windows PC, you can use the SSH Tectia Client that comes as part of the UWICK software set. If you need more detailed information on logging in, take the link on the StatGen computing web pages that says Introduction to Linux at Biostat (or this link here!).

  1. Students may prefer to use markov rather than abacus. Everything should work the same on markov as on abacus, but I will only use/check things for abacus. These are both 32-bit machines, although the machine markov is much faster. The other new Biostat linux machines gauss and hercules are 64-bit machines. Most things will work the same on them also, but not everything!

Now you are almost ready to start:


  1. To be able to use the StatGen programs, you need to fix up your .cshrc file to include the line

    2. source ~statgen/.statgen.cshrc

    You can do this using this command:

    echo "source ~statgen/.statgen.cshrc" >> .cshrc


  1. For step 5 to take effect, you will either need to log out and log in again, or else type the exact same line as you put into your .cshrc file at the command prompt. To check it has worked, type

% which pedcheck

The machine should with respond with "/home/guests/statgen/i386-linux/bin/pedcheck". It is telling you which pedcheck command it will use for you: pedcheck is a program we will use in the first lab. If the machine responds with "pedcheck: Command not found", something is not done right. Try again, and/or consult your peers. As a last resort, contact the Biostat computing staff, (bite@u).


  1. One last set-up thing that is not essential, but which you probably will find useful. Make a special directory in which to keep all your StatGen files this quarter. To make a new directory entitled “b550”, just type

% mkdir b550

Then, to get into your new directory, type

% cd b550

For more information on these commands, or on using a Linux or UNIX system in general, take the link UNIX/Linux tutorial from the Introduction to Biostat Computing page. The web page you are going to has the address: http://depts.washington.edu/statgen/Computing/linux-tut.html