It’s interesting to observe how your study habits change with technology. I now find that I only pay attention to journals if I’m subscribed to their RSS feed in my newsreader. I know there are the BaRf feeds and I could rig up something via PubMed but surely all journals should have a TOC RSS feed by now. That “sign up for eTOCs” link just looks so antiquated.
What does it mean? Where does it come from? Ever wondered the same thing about guts for garters, Bob’s your uncle, singing fat ladies and hundreds of others? Look no further than the excellent World Wide Words.
There may not be much new here for the next 2-3 weeks. After almost 6 years at UNSW I’m moving on. My new position is at the University of Queensland in Brisbane, where I’ve signed up for 3 years as a computational biologist with these guys. It looks like a great research project in a good working environment.
Wish me luck and see you on the other side.
DGPI is a software tool (Java) for the prediction of GPI anchors (GPI = glycosylated phosphatidylinositol), which various organisms use to attach proteins to membranes.
Are there GPI anchors in Archaea? We’re not sure. You can download DGPI for free to run locally but like many packages, it accepts only one input sequence at a time. Can we install it and analyse one of our archaeal genomes, asks a member of my lab?
So I whipped up a quick and dirty Perl wrapper. Given a fasta file with multiple sequences it takes each one, runs it through DGPI, parses the output and prints out a nice CSV file for spreadsheet import. One problem is that DGPI output is rather verbose and variable – without knowing all the output variations, we could be missing something when we parse. But the main thing is to extract a simple “yes or no” for each sequence, which is easy enough.
It’s quick – churned through 2 493 sequences in about 40 minutes on my slow 2.66 GHz Celeron machine and the parsing seemed to work well. And it identified 128 candidates which after all is the goal of bioinformatics – reducing the data down to something that you can sift through. Many of them appear to be predicted membrane proteins, some with interesting roles, so should be good times ahead for the lab.
Being trained primarily in biology, quantum physics is not my strong point. I find it fascinating, I enjoy reading about it and I wish at some level there was an intuitive way for a non-specialist to understand it. The problem is that even practitioners in the field describe it as counter-intuitive.
Today I’m trying to get my head around a Nature article, Counterfactual quantum computation through quantum interrogation. The essence of it, summarised here, seems to be that due to the probabilistic nature of photons, you can get them to cause effects without them actually being there. In this case, obtaining a result from a quantum computer without running its program.
No, I don’t get it either, but it intrigues me. Any physicists out there care to profer a simple explanation?
I like the Winter Olympics. They bring back fond childhood memories of Sunday tea at my grandmother’s house during winter, when we used to watch a classic TV programme called Ski Sunday. One of those long-running BBC shows with a stirring theme tune.
It occurs to me though, that we watch winter sports mostly for the accidents. Ski jumpers plummeting to the ground, bob sleighs tipping over, skaters glaring at their partner for throwing them onto their arse. How fitting then that in Australia, the Olympics are shown on Channel 7, purveyors of sensationalism par excellence. Every crash, fall and slip lovingly replayed in slow motion over and over again for your enjoyment.
Very cool – attach sensors to the head of a seal (no harm done!), they dive down to 2 000 m and on surfacing, transmit oceanographic data via satellite to base. SEaOS site here. Link from the news article to the TOPP website which includes near real-time data from tagged species in the Pacific.
Sometimes I wish I’d been a marine biologist. I was told at school there were no jobs in it…
Sometimes you read a science news headline and your first reaction is “I don’t think so”. That was my response when I read DNA could predict your surname. On reading the article, it becomes clear that “DNA could predict your surname with extremely low accuracy”.
The principle is obvious enough. If you take DNA from Mr X, his father, grandfather and uncles, there’ll be a fair degree of similarity, perhaps enough to say “it’s from the family X”. Where it all falls down of course, as the article states further along, is when adoptions, infidelity, name changes and multiple founders enter the picture. Not to mention the complete lack of application to female suspects.
Note to self: when booting working cluster nodes, ensure PXE booting is not first BIOS option, otherwise SystemImager runs again and destroys your hard work…
Some hours later, node is back to normal. Should probably use SI to backup nodes once installed to your liking.