Biomarkers biomarkers everywhere
February 28, 2007
Just attended a conference on biomarkers. Rich fodder for blogging. Hopefully at least a couple of posts over the next 10 days.
Further Reading:
Reality lies somewhere between hype and the naysayers
Technorati Tags: Biomarkers, Drug Development
“Data is often better than you think”
February 25, 2007
Gapminder has previously been mentioned in these parts. The quote in the title comes from a presentation on Gapminder by Hans Rosling from the Karolinska Institutet. I strongly encourage readers of this blog to check the presentation out. It raises many important questions about data, design tools, and trends, all subjects of relevance to biologists. Especially check out the section around the 16:00 mark about visualization and search. The take home messages for me were - public data is available. How can we use the data? What can we learn from all that information? How can we change the way we extract information from data sources?
Source: Blogspotting
Further Reading
Things I noticed #21
Visualizing data
Technorati Tags: Gapminder, Hans Rosling, Data Visualization, Data
80 cores? We’re still figuring out 2
February 25, 2007
By now, I’m sure that most people are aware of Intel’s recent demo of an 80-core chip. Teraflop number crunching on a single chip. Just imagine running 100s of nanoseconds of molecular dynamics simulations on your hope computer while you edit video on it as well. Reality check #1. The chip is a research chip and the technologies won’t be available for another decade (although the chip industry has a habit of surprising us). Reality check #2. The programmability of multi-core chips is still a huge challenge, and no one is really taking advantage of current multi-core chips yet, or at least not too many are, so it’s a bit of a stretch to imagine the scenario I listed above happening anytime soon. It is a little sad that potentially disruptive advances in technology have not yet been disruptive in any true sense.
Further Reading
The problem is the software
Get yer cores here
Bandwidth as a limiting factor
The rise of wildly speculative execution
Technorati Tags: Intel, multicore chips, high performance computing
Searching biological information at NextBio
February 25, 2007
John Cook at the Seattle PI points us to an article about NextBio. I first heard about the company when I noticed a press release mentioning that Jane Su, whom I have met a couple of times in a past life, had joined NextBio as Director of Scientific Computing. Leroy Hood’s involvement was something I was not aware about till I read John’s piece.
So what is NextBio? The company is a “web-based scientific data search engine that offers instant access, search and collaboration across a vast repository of life sciences information”. In theory, this is one of reasons that I’ve argued for data standards and open access. That allows companies like NextBio to build value from all the data available and adding on additional information as well. At least on paper, the approach the company has taken seems to be viable. They seem to be taking a Web 2.0-type approach. The company has challenges, something that John points out using the example of Nervana. There is always the build vs. buy argument. A lot of companies will need some convincing to go for the SaaS model that NextBio offers. Those long sales cycles that doomed Nervana’s life science efforts will need to be addressed. Hopefully some day I will be able to dig deeper into how the search engine works, and how it empowers the researcher. A search engine targeted at informaticians and statisticians will not be viable in the long term. Biologists must be able to access the information and be significantly more efficient than they would be with existing resources for the kinds of service NextBio is offering to be successful. The part that intrigues me here is the content. For a company the ultimate value arises from being able to add to any knowledge base and use that additional information in future searches. I am not quite sure if that is possible from the product description. A lot of thought seems to have gone into the infrastructure, essential for a successful search engine (Google rules due to its infrastructure).
There are more fundamental questions. What are the best search strategies for life science? What is the right approach to monetizing biological search? Can the NextBio model succeed or are there others out there, which make more sense? What do companies really need? What are the needs of the individual researcher? Biology is a field crying out for good search strategies search engines. No one has really succeeded in proper commercialization (not that I know off anyway). Will NextBio be the first?
Further Reading
It’s not just the EBI
Biology, search and Udell
Biological content: Access and monetization
Technorati Tags: Search, Search Engine, NextBio, Content, Bioinformatics, BioIT
On the road again
February 23, 2007
On the road this week, hence the lack of action (I’m not the only one). Might get a post or two in over the weekend and perhaps a couple next week. In the meantime, I highly recommend this interesting read on tapeworms
