Microsoft has signalled how seriously it regards the life sciences market with the launch, on 30 April, of its Amalga Life Sciences 2009 software. The development of this knowledge management platform complements an earlier initiative intended to foster a collaborative approach in this sector – Microsoft’s BioIT Alliance, which has been in existence for three years now.
Amalga uses semantic processing and reasoning technologies with the ultimate aim of providing personalised healthcare, through the provision of timely, contextual knowledge. The BioIT Alliance is a partnership programme whose stated aim is to realise the potential of personalised medicine. It has now attracted around 100 companies as members, operating in fields broadly applicable to data and knowledge management for the life sciences. Amalga software could speed up innovation within member companies.
Microsoft’s own Health Solutions Group has recently been expanding in budget and numbers of staff. Jim Karkanias, senior director of applied research and technology at the Group, told Scientific Computing World: ‘Life sciences, and the healthcare derived from them, can be viewed as an information-management problem: it's about finding the right piece of information, to deliver to the right person, in the right way, at the right time, and in the right context. Currently this process goes on in slow motion; researchers discover something, and it then winds its way through a fairly manual system of important processes, until it eventually shows up in the physician's toolkit. Even then it [may be some time before] an interested person finds out about its relevance.’
The hope is that Amalga will accelerate this process. ‘Once you have an information platform of this type, it can sort-out and separate facts before making their relevance known to the various interested parties,’ says Karakanias. ‘You can imagine a situation in which a scientist's work - relationships between an experiment and its findings - is automatically correlated to the literature and body of knowledge that already exists. That's what happens today - that is science - but it's a fairly manual process of flicking through journals, reading them, and understanding them. However, the facts and figures can now be connected by a machine, thereby accelerating that part of the process. If we can do that, we can connect the data to the physician; we could let him know that the presence of a given polymorphism of an enzyme has implications on how a particular drug is eliminated, and that it therefore has implications on the dosage that should be prescribed to a particular patient. If we can inform the physician, we can also inform the patient: "if you're taking this drug and have this particular metabolic pathway, (which, based on your genetics, you have) then your dose should be this. By the way, we noticed that your dose is not this."'
One of the main difficulties is how to store the data in such a way that a machine-like process can reason across all of it. Karkanias says that relational databases work up to a point, but it has been necessary move beyond them, towards a 'data net'. 'We view relational data as simply a special case of that broader way of organising data. The data net approach has allowed us to create reasoning environments on top of the data, to connect facts by navigating the graphs of data,' says Karakanias.
Karakanias concedes that: 'There exist a lot of technical approaches to the problem within the healthcare market, but so far they've all been just paving over the cowpats, and haven't changed the way things work.’ Microsoft believes that its BioIT Alliance and the Amalga could force the pace of change.
