Issue Date: May 28, 2012
Up From Chaos
Consider this definition of the word “genius”: Someone who extracts clarity from chaos and effectively communicates a meaningful vision to the public at large. Were we to accept this definition, it would serve us well to study the second step in the process. Communication may be what separates the true genius from the mere high scorer on standardized tests.
This idea of genius has some currency in the realm of medicine, where the scientific enterprise has long applied itself to navigating the chaos associated with human disease. In the process, however, science has added layers of human-made chaos associated with research tools and methodologies to the innate chaos found in nature. Although tools and methodologies can lead to cures, they also can create confusion and too often distract medical researchers from their primary focus, the patient.
There is no question that the explosion of heretofore inaccessible data on the nature of disease, generated by genomics and channeled through bioinformatics, has moved the ball forward in drug research. Access to reams of new data has also moved the game onto a new field, one that reveals the heterogeneous nature of disease. Genomic information has become the key to identifying patients who will respond to therapies that would not come anywhere near the clinic in a blockbuster drug development mode.
The annual Bio-IT World conference, which took place in Boston last month, provided an excellent window on the benefits as well as the distractions of the current data storm in pharmaceutical labs. The preponderance of data in drug research was made palpable with three days of 12 parallel sessions bearing titles like “An Algorithm for Identifying Multiply Modified Endogenous Proteins Using Both Full-Scan & High-Resolution Tandem Mass Spectrometric Data.”
Then, at a day-three panel discussion on challenges in cancer research, there came a moment of epiphany. Julian Adams, president of Infinity Pharmaceuticals; John Quackenbush, a bioinformaticist at Dana-Farber Cancer Institute and Harvard School of Public Health; and other panelists seemed to raise the collective call: “It’s the patient, stupid!”
The panelists discussed new approaches to making sense of the data generated in genomics-based drug discovery and development that have less to do with computers than with patients in clinics. The group envisioned a kind of virtuous circle in which data are put to the task of identifying patients most likely to respond to therapies, and trials are designed to vet the data coming in, eliminating the extraneous and creating knowledge applicable to clinical trials.
The current glut of data, however, can lay the virtuous circle somewhat flat. “For a lot of people, data do become a distraction,” Quackenbush, who is helping develop a master’s degree program in bioinformatics at Harvard, told me after the panel discussion. “People substitute our ability, now, to generate large quantities of data for the need to intelligently design an experiment.” He suggested, however, that the participation of the physician-scientist in drug discovery helps to counteract the seduction of complex data: It is much easier to stay focused on the patient when the patient is in your office, he said.
And it is much easier to get rid of extraneous data in the clinic. Jonathan Rothberg, a pioneer in next-generation genome sequencing and founder of DNA sequencer supplier Ion Torrent, credits physician-scientists with making the practice of genomics manageable in recent years.
By concentrating on the approximately 400 genes that the research community is most familiar with—genes associated with the mutations in more than 2% of patients with particular cancers—screening becomes a far less haphazard enterprise than taking on an entire genome of 25,000 genes. And with the increased focus on relevant genes comes a feasible means of substituting a manageable screening set of hundreds of genes for a single, possibly misleading, biomarker in clinical trials.
Genomics has reached a tipping point, Rothberg insists. “It’s the opposite of what genomics has been for the last 20 years.” Researchers are no longer gathering overwhelming quantities of data. “Now, they are working with panels of 200 to 400 cancer genes, knowing exactly what these genes do.” And the work is being done in the clinic. “I have never seen so much focus on translational use of sequencing as I have in the last six months,” Rothberg says.
The fact that Harvard is now contemplating a formal curriculum for bioinformatics may be further indication that some kind of clarity is emerging from the data chaos. Bioinformatics is becoming a discipline that will deliver tools and information to researchers “in ways that make sense,” Quackenbush said at the Bio-IT World panel discussion. Ideally, data will be presented in a format that allows clinicians to edit rather than analyze genetic content, he said.
Quackenbush went on to affirm the crucial second step in the genius process: “And we need to communicate the mission of what you’re doing back to the public.”
Back, that is, to the patient.
Views expressed on this page are those of the author and not necessarily those of ACS.
- Chemical & Engineering News
- ISSN 0009-2347
- Copyright © American Chemical Society