Collaboration Through Visualization

By Nick Basta, Contributing Editor

No one doubts that better collaboration among departments within a pharmaceutical company, or between it and its business partners, is a good thing. The challenge, however, is to collaborate effectively, without creating more bureaucracy and endless meetings.

There are two parts to solving this problem: having the right software tools to generate and organize information, and having the right work processes in place to make the best use of that information. A common feature of the latest software in planning and design is visualization, allowing users to see plant layouts, process flows, even the data involved in molecular design. The best work processes avoid repetitive data entry, requiring reliable ways to store and retrieve data, and allow it to flow seamlessly from one application to another.

Collaborate or Wither

How important is collaboration? "The best companies in our industry are dramatically increasing the collaborations within the different parts of R&D, and dramatically increasing the collaboration between R&D, [manufacturing] and the supply chain" said Fred Hassan, CEO of Schering Plough (Kenilworth, N.J.) in an address at Princeton University last month.

As part of an overall revamping of Schering-Plough, Hassan, who came on as CEO last year, has a multipronged improvement program going on within the company. Collaboration activities are part of what he calls "customer-centered product flow," which is meant to connect marketplace drivers in the industry more closely with manufacturing and R&D. "Interactions between R&D, the commercial organization and manufacturing were relatively infrequent until a compound was moved into development,and even then, the interactions might have been rare," says Hassan.

This interface,where information is flowing between basic or clinical research and process design,is being addressed by software vendors on both sides of the divide. On the research side, companies like Accelrys (San Diego) are looking to use their molecular modeling programs and databases to address process considerations such as crystallization or formulation.

"The drug discovery process has two steps, from our perspective," says Keith Glassford, director, materials science marketing at Accelrys, "gene-to-lead, and lead-to-delivery." In the first step, Accelrys tools such as Cerius, a combinatorial-chemistry program, are used to help identify candidate compounds based on the reactivity and conformation of genetic or cellular targets, such as the surfaces of viruses. A biochemical insight into a new treatment results in a lead drug candidate which can then enter the rest of the drug discovery process.

In lead-to-delivery, researchers are considering how to produce commercial quantities of the candidate, what its toxicity or interactions are in the human or other organism, and how to formulate an actual pharmaceutical product. Accelrys' tools include C² MedChem Groupware for analyzing chemical structures; additional modules provide for modeling catalyzed reactions, visualization of the molecules or their cellular level, or "meso-scale" properties. Other programs, DPD and MesoDyn, analyze polymeric structures for drug delivery techniques.

According to Keith, several other benefits come out of Accelrys' tools for managing drug discovery. One is C² Polymorph, which can predict polymorphic forms of pharmaceutical crystals. This becomes important during the patent-application phase of drug discovery, when a lab wants to protect all the polymorphic forms of the compound under the patent. Finally, the MS Fast program, based on an Oracle database management system, allows labs to store previous experimental results, and to determine what battery of experiments needs to be performed to characterize a promising drug candidate. "Labs are often forced to repeat 20-30% of their experiments simply because necessary data wasn't properly recorded," notes Glassford. This kind of inefficiency is the opposite of an effective collaboration.

So far, Glassford notes, it is rare for Accelrys products to be used by process engineers; however, the programs' capabilities are advancing, and their value in the process design stage is increasing. "Leading-edge companies know that these tools can speed up the lead-to-delivery stage of development, and that can make a significant impact on the bottom line."

Accelrys provides fairly sophisticated tools for visualization of candidate molecules, and their reactions with catalysts or with biological surfaces or molecules. "It's a cliche, but a picture is worth 1,000 words to show stereochemical effects, and to communicate results between research, engineering and other groups," says Osman Gumar, executive director of Accelrys' cheminformatics and rational drug design group. "For process design considerations, our software can do things like visualize surface interactions in gases or on metal surfaces, which could affect materials selection."

Visualize Design

When it comes to actual process design, being able to visualize the process is valuable, both in collaborative work and in meeting regulatory approval. Michael Grund, head of process engineering at Merck kGa (Darmstadt, Germany), uses the graphical and numeric output of a simulator to document a multipurpose batch plant, the largest in Europe, which has been built there over the past several years. "This has been a very complex process, with literally hundreds of different batch recipes needed to be reduced to a set of what we called 'design recipes,'" says Grund. "The documentation produced by the simulator helped us communicate with the project contractor, as well as to meet regulatory approvals."

Merck uses BatchPlus, from Aspen Technology (Cambridge, Mass.). Dan Mekler, business development manager, says that BatchPlus has utility at "both ends of the design process, either to work with research chemists to figure out the feasibility of a process, or to develop process flows and capacities, and from that the equipment specifications, with the project contractor." On the laboratory side, a user can deploy a 'route selection and screener' feature to evaluate recipe alternatives. On the process design side, a 'scale up algorithm' is used to raise the output of the process to kilogram levels or beyond.