In the next two weeks, my group at the University of Minnesota will be conducting a series of experiments in distributed computing. We hope to include a student lab of G5 systems running 10.4.x and Xgrid in this trial.

Specifically, we are looking for documentation and tools for running BLAST, preferably in a high-throughput capacity, on an Xgrid of 30 dual-processor G5 2GB/RAM systems in a local student lab. We will have cycle-stealing access during the day and dedicated access during the night.

I am familiar with the demo tools in the preview version of Xgrid, but they have vanished in the current incarnation. When we worked with them in the past we found them to be difficult to use and impossible for our users.

Optional queuing systems for this project include Sun’s GridEngine. We currently use SGE on three production clusters, but since this is a shared resource (a student access lab) the people who manage that resource are concerned about adding service software to these platforms.Xgrid side-steps this issue since it is already installed and requires configuration. (BLAST and associated targets of interest have yet to be installed.)

(update, 2006/08/16)

Xgrid is now enabled on our iNquiry cluster and I am meeting on Friday to enable systems in Walter Library. I hope to have some demo scripts in place before the roll-out. More about this when I have something to see.

I chose the Belkin Bluetooth 2.0+EDR dongle (F8T012v) with the Broadcom chipset. The box declares “Mac OS driver and software supplied by Apple”. I fully expected this device to work when I plugged it in. When it didn’t, I called Belkin. They told me “Mac OS driver and software supplied by Apple”. We opened a ticket and they agreed to escalate, all very pleasant but not very productive.

On a whim, I Googled my problem (‘Belkin F8T012v on a Macintosh headset profile‘). This should have been the tool of first resort because I found a link to a Belkin support site in the UK that provided a Macintosh driver, specifically.

After installing the Belkin driver for the Broadcom chipset on my OS X 10.4.6 PPC and rebooting, I was able to pair the 510 with the Belkin and make a clear and decent phone call via Skype.

I called Belkin, cancelled the open case and sent them the URL. Sigh.

I taught an introduction to programming workshop with an emphasis on Perl and extensions interesting to Biologists as part of the Summer Bioinformatics Institute at the Digital Technology Center, here at the University of Minnesota. Sixteen students, three half-days, labs, homework, etc. It was a blast. I thoroughly enjoyed myself and I think the students learned some useful bits. Kudos and thanks to my TA, Allison Langham.

Slides may contain images protected under copyright laws. If you have an interest in obtaining any portion of these presentations, please contact me directly. I am also willing to travel and train at your location.

(Handouts: Introduction to Programming Handouts, 28MB)

The recordings are offered in mp3 audio format and vary in length from about three to seven minutes. They can be found and downloaded here;

http://www.osa.umn.edu/podcasts/

This work received some mention in the local paper, the Star and Tribune.

Most of the recordings were created using Audacity and an Apple G4/Cube. Audio quality is lacking but better quality hardware has been purchased for future voice-over work.

I used ‘The Singularity is Near: when humans transcend biology‘, by Ray Kurzweil to support a Spring 2006 Bioinformatics Journal Club talk about the coming convergence of technologies in Life Sciences research. Specifically semantic web services.

In the book he states that humankind is at the threshold of an epoch and the rate of exponential growth is itself growing exponentially. ‘Singularity’ is described as the point at which a function takes an infinite value. In space-time, this is when matter becomes infinitely dense, as in a black hole and the normal Laws of Physics no-longer apply.

His point is two-fold:

• There are no constraints on human capacity, especially as we begin to build machines that learn from each other.
• This is a good thing.

How does this apply to Bioinformatics?

• Web interfaces, currently human-friendly, will become machine-friendly.
• Data formats and interfaces will begin to standardize.
• Heterogeneous platforms, applications, and systems will begin to interoperate.
• Machines will begin to communicate with each other in profound and powerful ways.

Why you should care.

• Distribution: data will continue to be created and controlled by autonomous groups all over the world.
• Biology is hard and messy: distributed resources and carefully negotiated ontologies will help.
• The key problems are social:
• We must learn to collaborate with service providers.
• By the way, you are a service provider.
• Technology is easy by comparison.

The current state of the art was punctuated with references to the Tower of Babel, by Pieter Brueghel the Elder, 1563.