The ability to do more in the laboratory with au- tomated solutions is fast becoming a reality for many in R&D. Automation, and the integration of other technological systems via Io T and analytics platforms like machine learning, is obviously going
to be a major part of the future of scientific laboratory
These technologies are already making waves in the
industry by removing mundane tasks from researchers
while increasing accuracy and efficiency. More importantly, it’s part of the “lab of the future” that’s no longer
in the future—it’s already here.
Automation, and other new technologies, influence the
way laboratories operate—how they collect the necessary
data to create drugs, form hypotheses or fund research.
With pressure to drive efficiency and cut costs, organizations are embracing automation—with impacts for both
laboratory scientists and the drug discovery processes.
Automation of routine “manual tasks” gets scientists
to sit up and take note. Robotics has been applied to
areas like high throughput screening in the past, but it
is becoming more pervasive now—making inroads into
DMPK, bioanalysis and process development. All can be
automated to a point, and this can have dramatic efficiency impacts, resulting in an increase of throughput. It
is a complex task and the “full automated lab” is a while
away, but progress is being made.
This is not simple and still requires detailed knowledge
of IT and robotics to get a solution working. Some instrument vendors are taking a more holistic approach to
the problem by merging many tests, IT and process into
a single instrument solution, thus removing the need for
bespoke software and robotics. This works extremely well
for known processes but is not so useful for adhoc or
emerging testing types.
Automation of “data discovery” is starting to emerge as
a trend with the advent of better storage infrastructures,
cloud availability and cheaper cost per Mb of data. This
coupled with rapid advance in automated analysis via
machine learning and deep learning tools is leading to a
seismic shift in what is expected of the scientist. We are
no longer accepting of the barriers to data analysis, and
this is likely due to the changes in the technology that
powers our daily lives. Everything is online—we have online banking, online insurance and online traffic/journey
routing, among others.
This transition offers providers with an opportunity to
The Automation of Data Discovery
By Paul Denny-Gouldson
Vice President, Business Development and Open Innovation, IDBS firstname.lastname@example.org
Editor’s Note: We invite you to submit your personal
commentaries to Last Word on topics that impact your
work and affect the overall industry. Please send your
commentaries to MaryBeth DiDonna, Editor, at marybeth.
differentiate themselves, making things easier with alerts
and suggestions such as “you look like you are travelling
to work—do you want me to give you the fastest route
based on current and historical traffic?” All these tools
are leveraging automated analysis and machine learning
to analyse people’s behaviours and get statistical models
for how and what people are going to do.
The critical part here is that AI and machine learning
etc are not “crazy new technologies”—they are all about
statistics and application of complex statistics to prob-
lems we see. What has made these technologies grow is
the availability of affordable, high performance compute
(HPC) power now associated with cloud infrastructures.
So, from a scientists’ perspective the future seems very
bright—and fast. What is still to be defined is how these
technologies can really impact scientists’ daily work, but
there are some exciting examples of what’s possible if we
look at the medical imaging world—especially automated
analysis of pathology image data. Here, the computer
analysis can increase the accuracy of the diagnosis by a
significant margin compared to human decisions. This is
exciting for us all. The key here is that the technology is
not replacing the human element but augmenting it. The
algorithm needs to be trained with the human knowledge,
and then updated with more human observations and
knowledge as time goes on.
From AI to robotics and from connected instruments
and virtual testing, scientists can expect to work at the
intersection of reality and science fiction. The imperative
to cut costs, increase efficiency and reduce time to bring
new medicines to market are forcing pharma and biotech companies to reconsider the ways they interact and
leverage technology. If organisations and institutions alter
their acceptance of new technology to reflect the rapid
advances being made, the reported stagnant R&D productivity could be halted or reversed. The incentives for
deployment of cloud-solutions, Io T, automation, AI and
other technology far outweigh the concerns—but choosing the right problem to solve, and with what technology,
remains the critical theme on everyone’s mind.