The Future of Structural Genomics?

I just came across this article in Science magazine that I believe is an interesting advancement in the field of structural genomics.
A group from Scripps and Burnham Institute including Adam Godzik took the whole genome of Thermotoga Maritima, a thermophile bacterium with a small genome, and modeled all the proteins and metabolic pathways, 478 proteins in total. This also included figuring out for a good number of proteins what the function of that protein is, and then reconstructing the metabolic pathways.
Of the 478 proteins, 120 had been identified experimentally. Of the missing 358 proteins, about half could be modeled with pretty good confidence (i.e. better than 30% homology). Only 3 of the proteins required some major tinkering to get at least a rough idea of what the fold looks like, and the analysis of the fold is what the group is focusing on.

First, I think the fact that two thirds of the structures were either from experiment or from reliable homology modeling is pretty encouraging, but it depends if you are a glass half full or half empty guy.

Second, once they had the enzymatic pathways modeled this way the group then identified a minimal set of proteins essential for the bacterium to survive. They found three groups of proteins: "core essential", where if you take out one, game over for Thermotoga, "synthetic lethal" where there is a built-in redundancy in the pathways such that one protein alone is not essential, but taking out more than one is lethal, and "non-essential" which are, well, non-essential.

The level of detail that the modeling allowed is pretty impressive (I think). I wonder how long it is going to take to go from this to a human or mammalian cell. An analysis like this could have a really big impact in drug discovery, where the preferred thinking of affecting one target to cure a disease is running into a bit of trouble lately, imho. Analyzing networks of proteins this way could really have a major impact on figuring out the best way to affect disease states using polypharmacology.

Of course, going from 478 to 20,000-25,000 proteins is not going to happen tomorrow (or ever?). Knowing how many proteins there are would be a good first step, I guess.
Did we finally figure out how many protein-coding sequences there are in the human genome? I have to check.

Ref.:
Zhang Y, Thiele I, Weekes D, Li Z, Jaroszewski L, Ginalski K, Deacon AM, Wooley J, Lesley SA, Wilson IA, Palsson B, Osterman A, & Godzik A (2009). Three-dimensional structural view of the central metabolic network of Thermotoga maritima. Science (New York, N.Y.), 325 (5947), 1544-9 PMID: 19762644

Towards an AIDS vaccine, once more

There is some hopeful news that an AIDS vaccine might be possible.

There have been multiple trials with AIDS vaccines in the past, all of which have failed, but now for the first time there is some clinical data that shows that infection rates were in fact lowered by the vaccination. Only by a third, though, so there is still a long way to go, but it is a step in the right direction.

Why, you ask, is an AIDS vaccine so hard to develop?

The reason is that almost all vaccines work on diseases that the human body can successfully fight on its own, mostly. Take for example smallpox, where about two thirds of the infected did recover from it before vaccines against it were developed. So all the vaccine needs to do is to prepare and boost the immune response ahead of time. This is how most vaccines work.

Chronic diseases on the other hand have managed to develop strategies to outsmart the human immune system. No matter how hard the immune system tries, for the vast majority of people the immune system can not clear the infection. So simply boosting and preparing the immune system through a vaccine does not work here, since there is no adequate immune response in the first place.

This is the reason why for this AIDS vaccine a different strategy was used. The vaccination consists of two vaccines, each of which works through a different mechanism. One of the two had been tested before and was found to not work well enough on its own.

The downside of this study is that the vaccines did specifically target HIV strains circulating in Thailand, where the study was performed, and even then only partial protection was achieved.

So maybe as for the AIDS drugs, the way forward for an AIDS vaccine is to have cocktails of different vaccines that complement each other.

If (big if) these initial results hold up.

Objects in the rear view mirror

There is an article in Businessweek about "How Science Can Create Millions of New Jobs". The article then goes on to lament how basic research exemplified by stalwarts like Xerox PARC and Bell Labs has declined and therefore needs a shot in the arm to create the jobs of the future.

I think the article misses the point by quite a distance. The question is not where did growth come from in the past and how we can revive these glorious times. The real question, I think, should be: Where is the current basic research going strong, inside and outside of the US, and in the areas where the US is lacking, what can be done to boost research?

The flaw the article makes is to only mention basic research in telecommunications and computer hardware, two areas which admittedly have driven job growth globally for more than a decade. But I think the cutting edge of this area and the big effect on employment is over. There is always room for surprises, but I think the fact that the likes of Xerox PARC and Bell Labs got downsized is not the cause of the drying up of basic research, but an effect of getting less and less money for a dollar spent on basic research in these areas.

So instead of looking at the past, lets move on, where is the new new stuff coming from? One way is to look at the areas where a lot of money is spend on basic research by industry and the public:

• Biotech, including bio-agriculture
• Medical research, for example analytics, medical devices
• Alternative/green energies, including things like battery technology, solar power, electric vehicles/hybrids, biofuels

In the first two areas the US is one of the major players, if not THE dominant player. But for a long time the US missed the boat and the potential for economic impact energy efficiency might bring. Americans insist on their god given right to drive oversized cars that consume bathtubs of gasoline, which is fine. But the rest of the world has gotten smart about energy efficiency and alternative energies. In the areas of hybrid cars, wind turbines, solar power, and biofuels the US is playing catch up with countries like Japan (no surprise here), but also Denmark, Brazil, and Israel, indicating that the playing field has been wide open for smart new entrants.

Fortunately the current administration seems to be willing to pour money into these areas now. The Xerox PARC and Bells Labs of the future are called DOE and Scripps (DARPA is still playing a big role).

The Businessweek article states "With upstream invention and discovery drying up, innovations capable of generating an industry have thinned to a trickle." I think the author is missing the torrent of innovations that have been happening while he was staring at his rear view mirror.

Research on the bleeding money edge

Here is an article from the Pittsburgh-Tribune Review about a lawsuit by a privately owned biotech in Seattle, Onconome, against Dr. Robert H. Getzenberg of Johns-Hopkins and U of Pittsburgh.

Getzenberg and Pittsburgh U hold patents for some cancer-related biomarkers stemming from Getzenbergs research. Onconome was funded to commercialize this finding and Getzenberg was the CSO of this company until 2008.

The lawsuit claims that Getzenberg made the whole thing up and the biomarkers never worked.

So yet another scientific fraud, maybe, but it brings up a good question: How do you find a balance between taking a wait-and-see approach on things, but risk that somebody else gets in early and reaps the rewards, versus get in early on some scientific discovery, with the risk that there might be some nasty surprises.

It doesn't have to be fraud, there are plenty of things where VC funds, biotechs, and pharma routinely spend a lot of money on things that do not stand up to the initial high expectations, aka hype (the human genome comes to mind). Some of these things do come back once the expectations are adjusted and turn out to be useful.

What I find mildly funny about the lawsuit is that Onconome is suing U of Pittsburgh for "failing to properly supervise Getzenberg's research". Never mind that Getzenberg worked for 7 years or so for Onconome and produced scientific results showing imaginary progress on commercializing the biomarkers, apparently without proper supervision as well.