Big-data is a perennial concern at Boston's annual Bio-IT World Expo because of the sheer volume of information the life sciences industry must contend with. The pain points expressed at last year's conference were twofold: big-data storage and big-data access. Now, the life sciences industry's concerns have evolved to those of big-data collaboration.
This month, at Bio-IT World 2013, speakers holding R&D and bioinformatics roles in the life sciences emphasized the need for externalization -- that is to say, collaboration with external entities, including competitors.
Externalization creates a bevy of technical needs that IT must address to allow better collaboration and data integration while keeping each party’s proprietary data secure. This includes having certain obvious capabilities in place -- such as videoconferencing, instant messaging, tightly controlled collaborative systems (e.g., Sharepoint), and desktop sharing -- and making them accessible to end users.
One of the more significant considerations, however, is one of big-data -- specifically, where do you put it all, and how do you manage it?
This naturally leads to the corollary question of whether or not to go to the cloud, and there is no single right answer. It all depends upon the particularized circumstances. Sergey Krymgold, associate director of Discovery and Preclinical IT at Biogen Idec, outlined for attendees several factors.
Legal and compliance issues comprise one factor, particularly when it comes to protecting intellectual property (especially from a competitor-collaborator). "A lot of times, people forget to ask, 'Who owns the data?' " observed Krymgold. With "first to file" patent protection (as opposed to the old system of "first to invent") having recently gone into effect in the US under the America Invents Act, Krymgold predicts that considerations of data ownership and data stewardship will play a larger role in collaboration agreements.
Data retention is a related matter. Because of intellectual property (IP) protection issues, a collaboration contract may provide that one or more parties need to erase certain data after the collaboration. This can therefore affect where and how the data should be stored. Similarly, regulated data, such as data governed by HIPAA and other laws and regulations, need to be given special consideration in determining how to store them versus their unregulated counterparts.
Other, "non-legal" distinctions of data type can also come into play. Raw instrument data may need to be stored differently than analyzed data and reports, according to Krymgold.
Krymgold went on to address how different data types may also need to be accessed differently. These and other accessibility factors -- including geography (such as performance at one access point versus performance at an access point on the other side of the globe), language requirements, and security -- will have impacts on IT solutions.
From there, Krymgold outlined the various pros and cons of internal and external hosting strategies, each pro and con taking on a different weight based upon the above considerations. A large company collaborating with a competitor may find more advantages to internal hosting, despite the initial investment, because of fewer IP security concerns, greater simplicity in following internal data retention policies, and ease of severing the relationship while protecting data. A small organization without an internal datacenter infrastructure, however, may benefit more from external hosting because of the smaller capital investment required and the ability to avoid potential integration difficulties with partners.
Either way, Krymgold and other panelists urged attendees to set standards (as well as support models), establish thorough business processes for externalized projects, train partners on both the applicable technology as well as one's own policies, and invest resources appropriately -- because the need for more, and more accessible, externalization is growing.
Frederic Bost, director of product management at Accelrys, highlighted this urgency with statistics. According to Bost, whereas 74 percent of projects necessarily involve two or more contract research organizations, only two of the top 20 pharmaceutical companies have a comprehensive externalization data management strategy.
"It’s different now," related Robert Boland, directing senior manager of External Innovation R&D IT for Janssen Pharmaceuticals, discussing external collaborations. "We were… having our door knocked on by small companies [to collaborate]. Now, we're knocking on their door."