What are the biggest scientific hurdles that the pharmaceutical industry is looking to solve with informatics? The short list may surprise you.
According to attendees of the inaugural meeting of Accelrys’ newly launched science council — which includes representatives from a baker’s dozen of the software company’s largest pharma customers — five topics stand out as the prize problems: ADME/tox prediction; ligand/protein docking; integrating clinical data into discovery research; workflow automation; and software-enabled medicinal chemistry.
Participants selected these topics out of around 17 proposed subjects that ranged from desktop bioinformatics tools, to protein informatics, to force fields for GPCR modeling, according to Scott Kahn, CSO at Accelrys. The final five topics — weighted far more heavily toward the chemistry end of the pipeline than the biology end — are in line with pharma’s desire to channel its expanding glut of genomics-based target data downstream, Kahn said. The mix of subjects “reflects how people are investing right now in bioinformatics vs. how people are grappling with the integration of genomic-based information into chemistry,” he said. “That’s really been a driver for us in providing some of that integrating technology.”
Kahn, who began planning the council over the summer, said the company would use the information gathered from the brain-picking session to guide its R&D process. Unlike the company’s domain-specific consortia, such as its functional proteomics consortium (http://www.accelrys.com/consortia/fg/index.html), where members pay a fee to meet regularly and exchange ideas that ultimately get incorpor-ated into customized versions of Accelrys software products, the science council has more of a longer-range charter. Kahn said that ideas gathered from the meeting would be used to steer the company’s long-term research investments. “We want to make sure that what we invest in is the right stuff,” he said.
At the first council meeting, held in view of the Manhattan skyline in Jersey City, NJ, on Sept. 5, representatives from companies like GlaxoSmithKline, AstraZeneca, Johnson & Johnson, Biogen, and Merck took part in a mix of short presentations from Accelrys scientists and council members, with a healthy dose of discussion, Kahn said.
According to Kahn, the meeting met its goals of identifying those scientific problems that are unsolved today, but could be tomorrow’s killer app. “There were two or three really big surprises,” he said. One, he noted, was “how virtual screening, in silico screening, needs to be considered from a research and development standpoint.” User objectives for virtual screening are different, he explained, depending on whether researchers are using the technology in early-stage or late-stage discovery. In early-stage virtual screening, researchers generally take a chemical genomics approach, using known compounds to gather information about a less well-characterized target. Here, Kahn said, the process is a bit more “tolerant of misinformation” than later-stage virtual screening, which focuses on optimizing a compound and demands a much higher degree of accuracy.
Kahn noted that he’s also “intrigued about the whole issue around the use of data [that is] further downstream and more tied to biology; that there may be many steps to be taken that in some respects may be easier to take. So, instead of saying, ‘Let’s take clinical data,’ maybe there are some things that are much more tangible that could happen before that. You’re moving more toward biology, but in a more realistic manner.”
Kahn said that the day’s conversation also addressed the “struggle to get good enough data that is consistent” to support in vivo ADME/tox prediction.
Accelrys plans to host future science council meetings via conference calls and WebEx online meeting software. The next face-to-face science council meeting is slated for its March AccelrysWorld conference in San Diego.