Friday, October 20, 2006


Spent today re-familiarizing myself with the power combo of Bison/Flex. Maybe because I'm a masochist, maybe cause I wanted to prove that CS class in programming languages, where we learned how to design compilers, was useful.

Bison is based off of YACC, which stands for Yet Another Compiler Compiler. And flex is a little program that creates C programs that tokenizes text based on regular expressions. What that means is that you take a piece of text and scan it with flex, and it turns it into a series of identified tokens. i.e. "1 + 2 is good" can be turned into the token string Number(1) Operator(+) Number(2) Word(is) Word (good). Then you use Bison to parse that string of tokens according to a finite state machine.

You can find nice little example at

Anyway, I have a set of features derived from several different aspects of protein sequence/structure alignment. I want to apply machine learning to these features figure out if these alignment are any good. But I haven't quite settled on the set of features I want to use. So I'm writing these little parse that will take configuration lines, like "%feature_1 / %feature_15 + %feature_2", and run through the database of samples, and produce the training files I will need for the machine learning.
Plus, I figure if I write this well enough, I can use it again later for other stuff.

Friday, August 11, 2006

Capsaicin the key to any wonder food

Capsaicin, the key ingredient in pepper spray and spicy food, has been found to cause apoptosis in 80% of human prostate cancer cells. Capsaicin, aside from being key ingredient in good thai and mexican food, also inhibits the activity of NF-kappa beta, which is part of the apoptosis mechanism.
It has also been found that spicy foods may also fight obesity. Capsaicin has also been used to fight pain (although over use may cause nerve damage). Scientists have also found that red peppers may boost insulin control , although that isn't linked directly the Capsaicin itself.

It may be time to start eating more spicy food ;-)

Contagious Dog Cancer

Scientists at the University College London have found a strain of cancer in dogs that is able to pass from one dog to another. Sticker's sarcoma, a usually non fatal cancer, typically resides in the dogs' genital tracts, and thus doesn't have to travel far to infect another dog during sex (it can also be passed by biting and licking). The original source of this cancer colony may have lived over 1000 years ago. Unlike most cancers, the genetic stability of Sticker's sarcoma has been credited with it's evolutionary success in surviving for so long.
You can also read some more about it at Scienceblogs

It's just weird to think about. Over a thousand years ago, some dog was born, and even today it's genetic code is still running around. Just imagine cutting off the tip of your finger, but it didn't die, it just kept on living and growing. I don't know how many cancer cells there would typically be in an infected dog, but lets just say 10 gram of mass. Assuming the average mass of a dog 25kg, it would take 2500 infections to get the mass of the original dog. According to MapsoftheWorld , there are 61,080,000 dogs in the united states. Assuming an infection rate of 3% (random number, not based on fact), that would be the same mass as 732 copies of the original dog. And that's just dogs living right now in the United States.

Monday, August 07, 2006

Futures in Biotech

Nice to hear some intelligent discussion about Biotech in the blogosphere (I don't like that word). Leo Laporte and Marc Pelletier have a new blog at
Also with a new podcast. There are only two show so far, but the first was on protein folding. Nice to hear, considering my main research thrust right now is protein fold prediction by threading techniques. They spend time talking about the 'big picture' stuff, which you can forget about when you are down in the trenches trying to get stuff to work for your PhD research. I spend most of my days worrying about debugging bad code in my threading software, or how I can improve my technique to make my results look better. I never really spend any time thinking about how my research could ever be used by anybody.

Changing the Definition of Vaccine

Looks like researchers at Scripps Research Institute in La Jolla, California have tested a vaccine for obesity in rats. I find this research interesting not because of the fixing obesity standpoint, but because they are developing an antigen for a hormone, Ghrelin found naturally in the body.

Microbiology has been changing the meaning of the word 'vaccine'. Wikipedia defines vaccine as "A vaccine is an antigenic preparation used to produce active immunity to a disease, in order to prevent or ameliorate the effects of infection by any natural or "wild" strain of the organism." Now we are talking about using similar techniques to train the body to down regulate hormone levels. I'd be curious to see what other lines of research are following this kind of technique.

Friday, July 07, 2006

The mad cows are coming!!!!!

Looks like someone is closing in on a detection method for the silent phase of Creutzfeldt-Jakob disease (CJD, or mad cow disease).

I'm interested in this because I spent some time in Europe in the mid nineties (near the time of the British mad cow scare). Because of that fact, I'm actually not allowed to donate blood. There is no screening system for blood samples, so they just have to exclude sections of the populations that may have contracted the disease. Strangely this didn't actually become an issue till about 2002. I guess they must have issued some policy change at the time. I had given blood before, and it was never mentioned. Then one day, they told me I wasn't eligible to donate. Oh well, it's not the biggest thing in the world. I would like to be able to feel like I'm helping out now and then, but I guess there are other things that I can do. But I did like giving blood because it gave me a chance to work on my fear of needles. It's not a phobia, but like most people, I fear them more then the amount of pain the inflict warrants.
Plus there is the small part of me that is secretly afraid that I may have actually contracted it. Of course that's incredibly unlikely, but if the odds are good enough for the blood people who knows?

Anyway, it looks like the process for detection involves a technique called protein misfolding cyclic amplification (PMCA). Basically, a sample is introduced into an environment with a large concentration of normal copies of the protein. The misfolded prion will go to work on those, converting them and causing chaining of the molecules. To speed up the process they break apart the chain with ultrasound.

The protein most likely behind this is PrPsc (normal form PrPc), the 121-231 domain has been solved via NMR, and the PDB code is 1AG2. The same domain (114-234) was crystalized in complex with an antibody for PDB 1TPX

Some Googling also found some research published in Proteomics. Essentially, they are trying to establish a link between an increase of cystatin C with the CJD. This is important because cystatin C occurs at levels measurable by mass spectrometry. And another previously linked protein usability 14-3-3 , is being questioned.