Eigencluster: Market Analysis and Strategy
Brief market POV from the PricewaterhouseCoopers report...
The introduction of e-R&D is a critical step. In Silico technologies will enable drug manufactures to accelerate the selection process, reduce the cost of preclinical and clinical studies and increase their overall chances of success. We estimate that they could collectively save at least $200m and two to three years per drug.
Yet most pharmaceutical companies are ill equipped to make the transition - partly because their IT is under-funded and overworked. They are already grappling with various compliance issues, the new technologies involved in early research and the corresponding increase in the output of data. But if the industry is to exploit the real power of e-R&D, it must innovate new technologies, build networked organizations and harness it knowledge capital. It must reinvent the role of the IT function. Above all, it must jettison the old, empirical way of doing things for systemic, predictive processes based on a more complete understanding of how the human body works.
I have looked around a good bit today and there is much data on various firms selling software/information services in the life science field. My sense is that we really need to pick as specific of an area/application as possible to develop this business plan. Otherwise, we run the risk of committing a start up sin... being a technology in search of a market! If this happens then we'll go through the entire semester without any true grounding.
I might suggest the following "animal research sector" is one way to be true to the "clustering" methodology of Dr. Vempala. The core ability of the technology is to manage the rapid increase in the volume of readily accessible data and then to locate relevant information and organize it in an intelligible way. I don't think the following animal application is a big leap of faith.
Also consider this...
The ability to turn data into knowledge and to do so quickly is a key competitive advantage. The massive inflow of data to the pharma/biotech industry has filled the "fact gap" that once existed. Now the business is swamped with information. This problem is exacerbated by the fact that while some kinds of knowledge remains relevant for years, others have a very short shelf life. Information that could shape the process of identifying a potential lead compound is just such an example. The ability to use data promptly and effectively is vital.
So here is one specific example that we could focus on...
Here is the overview... In developing lifesaving drugs, animal research is necessary to confirm the safety, efficacy, timing and dosage for these new compounds. Unfortunately animal testing is a costly and an inexact science and currently 92% of the drugs that succeed in animal testing will fail before product launch.
A number of intrinsic problems in animal testing contribute to this high failure rate. Although relevant animal models are used, the results collected from animals do not directly extrapolate to humans. Also, the number of variables tested is inherently constrained due to time, resources and the physical requirements of the tests. Therefore the researcher must make many assumptions and educated guesses to design and develop an effective protocol.
Finally, scrutiny is increasing for the use of animal testing. Pressure exists to replace animal testing when appropriate and to collect more effective results while reducing the number of animals. This directly addresses the IACUC mandates for replacement, refinement and reduction in animal usage. (http://www.iacuc.org/)
Animal testing via an in silico model is an idea that is simple to understand, plays directly into the technology. There is tons of data for various animal trials and though not a fully baked idea this direction is worth talking about.