Workflow and Pipelining in Cheminformatics
Commercial considerations
As an over-generalization, one could say that SciTegic is supplying "vertical solutions", specializing in cheminformatics, while InforSense is spread across a group of diverse "horizontal" fields. InforSense does report many users in pharmaceutical companies, and ChemSense has the advantage of being vendor-agnostic, but Pipeline Pilot has made more impact with computational chemistry departments, partly, perhaps, because of the established reputation of Accelrys with those groups of users.
InforSense’s marketing pitches use terms such as "business intelligence", "embedding intelligence throughout the enterprise" and "pervasive, process driven and predictive analytics". The company is aiming not just at pharmaceutical research, but also at medical research centers, sales and marketing groups, and operational risk compliance programs. It has had large injections of vendor capital of late, allowing it to diversify in a big way. Its in-database technology is probably of more importance in some markets than in others. Pipeline Pilot does not have in-database technology, but SciTegic has pointed out that, in practice, in cheminformatics, researchers tend to have data in all sorts of files.
It is probably fair to say that Pipeline Pilot is the market leader in the cheminformatics niche (there are only two hits for InforSense in SciFinder but more than 20 for Pipeline Pilot). SciTegic does have customers who use the infrastructure and reporting for completely different tasks (for example, one company has built a CRM from it), and some customers in the imaging and bioinformatics areas are not in the field of cheminformatics. SciTegic and its customers, have, however, focused their scientific publications on cheminformatics. Recently, SciTegic like InforSense, has started to talk of business intelligence, at least in terms in terms of "Scientific Business Intelligence". Accelrys appears to have trademarked this seemingly generic term.
Some people have an impression that commercial solutions have focused on IT infrastructure, perhaps a good theme on which to focus commercially. InforSense emphasizes database technology. Accelrys’ own algorithms are delivered through Pipeline Pilot with Discovery Studio or Accord for Excel as the graphical user interface. Pipeline Pilot acts as the middle tier of Accelrys’ architecture, sitting on top of databases and computing servers with molecular modeling code. It is certainly true that many users consider the two commercial solutions expensive. The different licensing model of KNIME is thus attracting interest. Equally important is its reported user-friendliness.
Not only does KNIME allow "best of breed" rapid deployment of workflows but as the pharmaceutical industry moves to an environment of shared risk, outsourcing and collaboration then it needs to be using tools, models, and frameworks that are cheap and simple to deploy. KNIME provides a company with a workable solution to share its models with organizations that cannot afford a complete, commercial modeling suite.
KNIME bridges the gap between the expensive commercial solutions that big pharmaceutical companies can afford, and solutions such as Taverna at the other extreme. KNIME does have some free cheminformatics nodes such as CDK, but Taverna is truly free and its developers have a strong preference for the open source philosophy, although there is, in practice, no barrier to interoperability with closed source systems (such as in the system built with Schrödinger nodes). Vendors of cheminformatics components claim that there is little demand from their customers for wrapping the components up in Taverna implementations but there are certainly some users of Taverna in the pharmaceutical industry, perhaps without formal management backing. Changing requirements, such as big pharma’s need to share components and workflows with collaborating organizations, and the need to lower some of the barriers between bio- and cheminformatics applications, suggest that the workflow paradigm has an interesting future.
|
