Wolfram Finds Few Answers Searching His Personal Genome
By Kevin Davies
April 5, 2011 | As part of his long-standing interest in making the world’s knowledge computable, Stephen Wolfram is taking a long look at model genomes, not least his own.
Last year, the British science prodigy, CEO of Wolfram Research, and developer of Mathematica and Wolfram|Alpha, had his own genome sequenced by Illumina. Wolfram recently spoke to Bio-IT World and offered a preview of his keynote at Bio-IT World Expo on April 12.
“I’m always an early adopter of technology tools,” says Wolfram, who was one of the first recipients of the MacArthur “Genius” Award in 1981. “I was curious to understand something about the personal genomics area.” Wolfram remembers that he received his genome in “a nice box” that contained a tiny disk drive.
“I had someone spend several months trying to dig through it to find something interesting from it. I’m a little disappointed [in the results],” he says. “I’d done SNPs previously. I’d found that my eyes are brown… I found I should be lactose intolerant, which I’m not. Otherwise, I’m pretty average, which I might have guessed.”
Wolfram recalls his colleague telling him there were “770 diseases” that he might have. “My first point was, most of them would have killed me in the first week of life.” Wolfram says he is “eagerly awaiting a lot more data on what is the correlation between all this genomic information and what actually manifests itself in a clinical way.”
The flagship product of Wolfram Research is Mathematica, a system for computing complex mathematical problems, which was introduced in 1988. The software is used by a broad cross-section of R&D users across all industries. “We’ve developed the symbolic programming language to handle image processing, control theory, statistical things, and bioinformatics capabilities,” says Wolfram. But his major goal is “to let the humans decide what computation they want to do… and let Mathematica automate the selection of which of 300 algorithms is best for that sort of thing.”
More recently, Wolfram launched the Wolfram|Alpha search engine, or more appropriately, “answer engine.” The focus of Mathematica, he says, is building a systematic framework so knowledge can be implemented. For 30 years, Wolfram has pondered whether he could take knowledge and make it computable. Eventually, Wolfram realized that “it wasn’t totally crazy to build a systematic system that could take a large swath of the world’s knowledge and set it up so, if you could ask a specific question, you get an answer.”
Utilizing some 15 million lines of Mathematica code, Wolfram admits that Wolfram|Alpha sounds tremendously ambitious. The answer engine spans science, engineering, finance, socioeconomics, general interest, even culture. Ingesting the raw data is in a way the easy part. “[We] find definitive sources of data and make them computable,” says Wolfram, the key being to find the primary, definitive sources of data. Merely “foraging on the Web doesn’t work at all,” he says. “The other 95% of the work ends up taking that raw data, setting it up so it’s all validated, organized, and correlated, so you can answer questions based on the data.”
In the life sciences, Wolfram says the answer engine can return all sorts of information, from the melting temperature of specific oligonucleotide sequences to matches with the human and other genomes, as well as information on single nucleotide polymorphisms (SNPs) and gene homologues. There is also considerable information on public health and medical test data. Type in your cholesterol level, and it will tell you where you fall on the population distribution.
Above all, Wolfram stresses that “humans have to communicate with [the engine] and tell the system what they want to do.” A priority was to develop natural language methods for querying the search engine. “That takes breakthroughs in linguistic processing and a lot of work in every area,” says Wolfram, noting the linguistic curation of gene names and medical diseases. “In every sub-discipline, there will be particular slang ways that people refer to things. It’s an interesting challenge to understand the grammar of that slang and be able to implement it.”
“It’s a pivotal time for the computer industry right now,” says Wolfram. “Expectations are changing. We’re excited about this knowledge-based computing that’s been made possible by Wolfram|Alpha.” Among other things, Wolfram says he is interested in very large-scale systems modeling and the future of biomedicine. “With all the sensors we have available, we can get all these data on a particular person flowing into the system. Is there a giant dashboard for each person, where we say, can we tweak this particular piece and now we can immediately compute the consequence of eating more whatever? Or in three years, you’ll have this issue that will cause you to take this drug?”
Another intriguing idea is what Wolfram calls “algorithmic drugs.” “Right now, we make drugs with particular targets and purposes... Systems like molecules can do very non-trivial computations. When you have your drug going around [the body], can you have it not just respond in a very standardized way to a particular target, but have it compute what to do on the fly, so to speak?”
Evidently Wolfram will have plenty to discuss in his opening keynote on April 12. “I’m looking forward to it—it’s a good moment to take stock of things we’re doing technologically in the bio-IT direction. I’m going to have fun.”
A recording of the full interview with Stephen Wolfram can be found here: http://chi-imagehost.com/podcasts/StephenWolfram.mp3