Tuesday, September 28, 2004


At least now I have an explanation for my insomnia last week : I'm sick! Some sort of bad cold / RSV (Respiratory Syncitial Virus) infection. Did part of my CIHR doctoral award application yesterday at home, since I woke up at noon, half alive but feeling mostly dead. Feeling a little better today... hopefully at the end of the week I'll be top shape!

Working of a visual redesign of the site, cause this default theme generated by blogger is too commonly used (and honestly, I find it ugly!). Not much experience with CSS yet, but I suppose its a good time to learn!

Until I get better...

P.S. : Did not even get hallucinations past 4 days of insomnia. False publicity I say!

Tuesday, September 21, 2004


Been unable to sleep for 2 days straight, for no apparent reason. Well, I did fell asleep yesterday, but at 5 AM, and my clock rang as usual at 7 AM, so I'm not sure it counts.

When I get better sleep I'll post about problems encountered while doing microarray analysis, principaly caused by non-standard gene/protein nomenclature in publications. Basically, when a gene is discovered by many labs, each one insist to name it as they want, which can complicate research to a point you can't imagine (Record to beat, 10 totally unrelated names for a protein). When you throw splicing in... it get spicy (ok, I'm totally sleep deprivated).

Saturday, September 18, 2004

Definition of bioinformatics

Had an interesting discussion about the definition of bioinformatics. A senior bioinformaticist (thanks kirwilliam! :)) at my University was on the committee to approve my doctorate program. For him, a PhD in bioinformatics should ONLY be involved in the programming of new software. I believe in (and submitted) a more balanced approach, with both wetlab and software development. The wetlab part involve almost exclusively microarrays (and their subsequent analysis, which as you may know involve PLENTY of bioinformatics tools). The software development part will improve and automate current analysis techniques of microarrays (can't really say more at the moment, except if you want to participate! :)). To make a long story short, we argued about it for about an hour, and my PhD application was accepted (they changed only the program's title, from "bioinformatics" to "bioinformatics and retrovirology") ... but this made my realize that the definition of the bioinformatics field, and thus what is a bioinformaticist (and what he's supposed to do) is still somewhat vague, leaving place to interpretation. I'll say it again, in my opinion a balanced approach is to be privileged : it helps you understand biological problems better.

As a side note, this blog got past the 500 hits mark, which I'm kinda proud of (I know, it's nothing, but it's my first website ever, so everything counts ;)). If you have suggestions of things I could add / change, please let me know. In the future, I plan to write short reviews of bioinformatics software (commercial or not) I use, random thoughts, plans to save the world, etc... :)

Sunday, September 12, 2004

Bioinformatics, or Will the bubble burst before the boom?

I was visiting some bioinformatics related forums this week, and what I saw frightened me. A lot. Remember the good old days of the dot-com era, where you could IPO for millions presenting the vague idea of a project? Lots of my friends studied in informatics back then, even if they didn't have any affinity or interest with the field, believing in the promise of a '100% garanteed job, we NEED informatics guys, it's the NEXT BIG THING'. Where are they now, in the post-dotcom era? Working at Walmart, working in doughnuts shops, etc.

What I saw on some bioinformatics forums made me believe that the same exact thing will happen (minus the massive IPO craze) with bioinformatics. Hordes of people asking advice to get in the field, obviously because it's the 'NEXT BIG THING'. I saw physicists and mathematicians with no programming knowledge nor biology insight get in the field. Hordes of Indians too, probably with some informatics degree, wanting to specialize themselves in bioinformatics (no problem with that), but having no interest whatsoever in biology. One even asked if it was possible to do bioinformatics 'without hardcore biology'... Ugh.

News flash : The bubble will burst, cause the world don't need billions of 'bioinformatics guys' who 'studied' in the field without having affinity or interest in it, hoping for an easy 100K+ job. Survival of the fittest will prevail, as ever. Interested in bioinformatics? Do learn programming, but do some 'wetlab' biology if possible. You can't understand problems biology has if you never worked with biological systems!

Thursday, September 09, 2004

Did I praised Adsense earlier?

Well, I don't say Adsense is bad... far from that. I spoke about AIDS for two posts, and got ads about... AIDS. But some were about... 'natural chinese ways to cure AIDS, stop using drugs!' and the like. I DO NOT approve any 'natural' or 'homeopathic' treatment of any sort for any disease. I banned these URLs from being displayed by Adsense... if you spot more, please warn me as soon as possible.

Historical lessons 1 : Homeopathy draw its roots (well, at least its 'credibility') from an article published in the highly prestigious journal Nature :

DAVENAS E, BEAUVAIS F, AMARA J, OBERBAUM M, ROBINZON B, MIADONNA A, TEDESCHI A, POMERANZ B., FORTNER P, BELON P, SAINTE-LAUDY J, POITEVIN B, BENVENISTE J: Human basophil degranulation triggered by very dilute antiserum against IgE. NATURE, 1988, 333-816-818.

This paper basically said that water kept 'memory' of molecules that it once contained, so that even by diluting a concentrated solution so that no molecules are present (theorically), water would preserve the molecules function. It got published. In Nature, no less. The guy in question, Benveniste, even won the Ig Nobel prize years later for this and unrelated work, adding to the 'confidence' general people attribute to his 'work' (see the confusion it cause? Ig Nobel Prizes are awarded to worthless science that couldn't be reproduced). However, the Nature journal sent a team to his lab to confirm his observations, double check his protocols, etc (this is a condition you must accept to publish in Nature). To make a long story short, results couldn't be reproduced by individual labs, and it made a BIG scandal in the scientific world. Benveniste lost his funds, his lab, etc.

What's sad it that homeopathethic (heh) people picked this 'proof' to 'confirm' that the FUD they're selling at 40$ a bottle really works. They say... see? It's proven! Nature! Nobel Prize! Biggest Charlatans (is this even english?) of the modern era, abusing people's confidence.

Even sadder : lots of people actually BELIEVE and DEFEND the guy (Benveniste). Google for his name just for fun; top 20 pages are pro homeopathy FUD.

The saddest part is Benveniste himself, who now works on (hold your breath) 'digital biology'. What's that you say? The ability of (quote from his bullshit website "believe that the specific activity of biologically-active molecules (e.g. histamine, caffeine, nicotine, adrenalin), not to mention the immunological signature of a virus or bacterium can be recorded and digitized using a computer sound card, just like an ordinary sound". Wow. He even say it can be transmitted over phone or the internet and still 'transmit' the biological function of the molecule. The guy clearly lost it... Seriously.

Wednesday, September 08, 2004

Allegory explained

First, the city is - you guessed right - the body. The citizens are cells; varied in form and function, working in buildings (organs, if you wish), traveling through a complex road system (the veins). Bacteria, and viruses (Bugs and vermin) are the unwanted strangers; the first category can multiply in 'buildings' (organs), using resources and causing damage. The second category, viruses, must infect cells in order to complete its life cycle.

Fortunately, the immune system is there to control infections. The 'death squads' in the allegory are CD8+ T lymphocytes and NK (natural killer) cells; they recognize infected cells and kill them, limiting the replication of unwanted viruses (or parasites). Decontamination teams are B cells, which are specialized in antibody production. Antibodies are protein that can recognize very specifically any 3D structure formed by molecules. Once coated by antibodies, bacteria/viruses get eaten by 'cleaning teams', macrophages and dendritic cells. The job of these cells is to pick up proteins in their immediate environment, process (digest) them and present fragments to the 'officers', CD4+ T lymphocytes. After a CD4+ recognizes something foreign, it is activated, multiply and organize the immune response against the 'stranger'. Of course, this is a much simplified overview of the immune system, but for now it'll do :)

HIV-1 can infect both CD4+ T lymphocytes and macrophages. These cell types harbor the HIV-1 receptor, CD4, and coreceptors (either CXCR4 or CCR5, chemokine receptors), which interact with gp120, the envelop protein of the virus. Of note, the simple interaction of gp120 with CD4 (in fact, multiple interactions are necessary, there's a certain threshold), is known to cause apoptosis (programmed cell death, or suicide) of CD4+ T lymphocyte; it is believed that it's the principal cause of CD4+ loss associated with AIDS progression. The loss of ‘officers’ impede the normal immune response against other pathogens; nobody dies of AIDS, but of the subsequent uncontrolled infections. HIV-1 is part of the retrovirus family; thus, it integrates its genetic material in the host's genome upon successful infection. Then, it can either go in ‘massive production’ mode, or ‘total invisibility’ mode. The later is the source of HIV-1 resilience. Government weapons are drugs designed by pharmaceuticals targeting various aspects of the virus life cycle; retrotranscription, maturation mediated by the viral protease, and entry. New viruses can’t reinfect new cells when these drugs are present, and infected cells in the ‘mass production’ mode don’t last long (few days for CD4+ T cells to a week or more for macrophages). The problem lies in the ‘dormant’ HIV-1; cells that harbor it don’t die quickly (recent estimations, half-life of 6 months or more, which mean total elimination time of > 70 years, assuming no new reinfection). Drugs have horrid side effects on the long term; patients can’t be on these for their lifetime non-stop. So they have ‘programmed interruptions’ where they stop taking the drugs for 6 months (give or take). In this period, a fraction of the dormant viruses wake up and reinfect the whole system anew. HIV-1 mutates at an alarming rate (2-5 mutations per new infection, estimated); it develop resistance to drugs VERY quickly. Thus, drugs are used in combination to retard resistance, but eventually (with the help of programmed interruptions, or non adherence to the strict multipills regimen two or three times daily) it develops. As we speak, there is NO method to eliminate HIV-1 once it infected successfully.

Saturday, September 04, 2004

HIV-1 allegory, take II

Ok, I'll try this again... hoping my browser won't eat my post before I'm finished!

Imagine a huge city, buzzing with millions of different people. It's very active, growing steadily, exchanging goods with it's neighbors. Unfortunately, it is periodically infested with bugs and vermin, which can harass citizens, contaminate buildings, etc. If uncontrolled, they can wreak havoc on the whole city, potentially destroying it.

Don't worry too much! There's an elaborate system whose only purpose is to control these infestations. It is composed of four (to simplify a little) specialized agents. First, there's multiple squads specialized in the hunting of bugs; each is highly trained to search and destroy one specific class of bugs. Second, there are decontamination teams; they research bug-specific sprays and disperse them on a city-wide scale, who recognize and incapacitate unwanted 'strangers'. Both of these are under the control of officers, which, after they're alerted of infestations, alert both teams, 'activate' them, causing them to multiply and overwhelm the enemy. Last but not least, cleaning teams pick up corpses and various debris and alert officers when they find something unusual. Everything's working perfect for most bugs; they get spotted, targeted and killed, causing minimal damage.

However, a new kind of vermin, called HIV-1, is uncontrollable. It target specifically officers and cleaning teams. It can harass officers to the point where they cause suicide, or infect them, become a part of them(few other bugs can do that; this particularity make HIV-1 invisible to the system while it's dormant) and either produce new bugs, quickly killing it's host in the process, or go in a latency mode, where it's undetectable. It can also infect cleaning teams in the same way, except that they last a lot longer, and thus produce in the long time more bugs. Officers and cleaning teams get replaced, but in the first case, the supply of officers is not quite enough to replace losses. Their number slowly but steadily decrease up to the point where there's not enough officers to control others infections. At this moment, the city is overwhelmed by bugs of multiple kinds and sources, and is eventually destroyed. What's worst is that HIV-1 is easily transmissible from city to city by certain types of exchange channels; it propagate rapidly throughout the world. The government tried to develop weapons against it. They are effective at stopping HIV-1 from infecting new hosts; the ones that are infected and producing bugs eventually die, like we said. However, the major problem is in infected cells where HIV-1 is dormant... it stays there until the army is gone (the army can't stay forever, citizens are not happy under martial law) and they reactivate and reinfect the whole city. To make things even worse, it mutate at an amazing rate, allowing it to evade decontaminating teams, death squads and government weapons. The last category must now be used in combinations, to retard the evolution of weapons-resistant bugs. No other strategy have proven efficient enough.

There you have it! Next time I'll post the real world equivalent, so you can make links and understand things better... :)

Wednesday, September 01, 2004


Almost finished an elaborate post where I explained HIV-1 with an nice metaphor... and I hit the back button, browser eats it, and it's almost midnight :(. Will have to wait tomorrow at least :) Meanwhile, you can notice that Adsense now allow us to place up to 3 ads per page, which is nice and should increase revenues (which are right now abysmal, but it's ok, just starting). I opted for 2, a top banner and a side one, at the bottom... the ads are very relevant AND interesting, to the point where they interest me, but I can't click on them (by contract with google, to avoid fraud) so I have to write them down manually in my address bar... how ironic :) Hope you like them too!

Got interested in microRNA this week... excellent primer by the mean of very complete reviews in Current Opinions in Cell Biology. Basically, it's a whole new field of gene regulation to study. Very small RNA duplex are produced by a complex machinery (involving Drusha and Dicer; the first cleave the precursor (pri-miRNA) at stem-loop structures to form 70 nts molecules (pre-miRNA), then Dicer cleave those in 22 nts fragments (miRNA, also called siRNA). Then, 2 complexes use these fragments to silence gene expression : RISC, which cleave mRNAs and RITS, who silence expression at the chromatin level. Very interesting (and new!) stuff... of course, siRNA (synthetic miRNA which are injected in cells to knockdown gene expression) is a big application of this discovery. We'll see if we can find a new strategy using them to cure HIV-1 (Many already did proofs of concept; few examples here, here and here)

As a last note, thanks to the good soul who provided me some Gmail love!