Expert view: Improving R&D productivity with novel screening approaches
Posted: 4 June 2018 | Steven G Jarvis (Head of Pharma & Biotech Segment Strategy - MilliporeSigma) | No comments yet
The pursuit of improved R&D productivity rates coupled with decreased cycle times has dominated pharmaceutical discourse in recent years.
As this trend continues, the shift to more novel screening approaches will continue to be prevalent in the discussion.
To achieve these goals, the pharmaceutical industry has shifted to employing novel strategies, such as shared-risk, by partnering with academic institutions, spinning out biotech start-ups focused on orphan diseases, and engaging in pre-competitive discussions with current rivals. While many of these approaches require engaging outside any lab in question, other methods to keep it ‘in-house’ will continue to be necessary.
In order to mitigate risk and ensure confidentiality of their data, pharma will start / continue to employ methods like artificial intelligence and big data on their path to in silico screening. Artificial Intelligence (AI) systems are becoming more complex and have been shown to produce outputs with much greater biological significance and accuracy – potentially leading to the identification of molecules that are highly selective and efficacious.
As the AI space continues to advance, one must recognise the major limitation – the data output is only as good as the data input. AI can be limited inasmuch that it is not able to develop novel screening approaches and thus needs new physical screening methods with robust data that can be assimilated and then disseminated into newer outputs. The data fed into the programme is absolutely integral in developing translatable in vivo results.
A newer screening approach such as pathway disruption through a proteolysis targeting chimera (PROTAC), which induces selective intracellular proteolysis, can be used as an example. An annotated platform combined with robust biologically relevant data would be fed into an AI to then generate predictive and accurate in silico results, creating promising paths forward to the clinic.
Using this approach, more data will be available about candidates during the screening process, leading to higher quality NMEs moving to the pre-clinical phase and this, in turn, will result in improved R&D productivity and decreased cycle times.