There is comfort in habits. There is
comfort in the familiar. In normal times, change usually only occurs at a
relatively slow pace. When we update one of our cell phone applications, it
usually takes a while before we adopt this new fancy functionality and take
advantage of it. After all, it has been working well for us even before the update,
right? So why change if things are fine? Even in our work, we find comfort in
routines and repeating tasks that we feel we are good at. But there are changes
that we cannot hide from.
Inevitable change
Everything around us is becoming “digital,” and the COVID-19 pandemic has dramatically accelerated the phenomenon. As a consequence, we cannot avoid upskilling to become digitally proficient, and we must do this across every sector. As pointed out by McKinsey, “prior to COVID-19, companies were experimenting with new technologies and finding it challenging to encourage software adoption amongst employees”.1 This was also acknowledged by pharmaceutical companies as they tried to roll out data science tools within their drug discovery workforce. Lewis et al. indicated that there is still much to learn to improve the tools and support the discovery chemistry workforce in the adoption of these tools.2
Data science (and related AI/ML efforts) will undoubtedly modify the way chemists work, but in order to capitalize on these endeavors as soon as possible, data science skills need to be infused into chemists’ training. Together with data science skills, recognizing employees’ fear of being replaced, leveraging data savvy workers and allowing people to gradually change the way they address daily problems is primordial. You can then hope that more chemists adopt tools developed by experts (data scientists, computational chemists, etc), but also that their feedback to these experts become more and more precise as they better seize the way these tools works and support them.
Upskill in pandemic time
Working closely with scientists from big pharmaceutical
companies in 2020, Elsevier participated in the accelerated digital
transformation and upskilling of these companies’ drug discovery workforce. As
the year lingered on and most of the scientists had to work remotely, our
discussions with scientists moved away from “how to retrieve an article” to big
data approaches and how to merge and analyse datasets for greater insights. We
focused our engagements on how to use the data available to them, on cheminformatics
skillsets and basic concepts used in data science.
Away from the lab, several scientists responded by switching their “microscope” to a “magnifying lens” and started analyzing both internal and public data from a larger perspective. Elsevier capabilities for integrating various chemistry sources played a role in the democratization of this type of analysis within these organizations. Indeed, by having their internal reaction data and the public reaction data available on the same platform, Reaxys, chemists usually not involved in big data approaches started to own such initiatives. Working closely with Elsevier consultants to understand the data structure and its complexity, the data was accessed via API (Application Programming Interface) and fuelled user-friendly tools for data analysis such as KNIME by scientists of all backgrounds (learn more about Reaxys API here). Projects around predictive toxicology, reaction prediction, reaction data cleaning and druglike fragment analysis arose from our discussions.
Along the way, scientists from these big
pharma organizations rallied around their in-house data science experts and
Elsevier consultants enabled these multidisciplinary teams to challenge each
other and learn from one another. The enthusiasm that emerged in this peculiar
year will undoubtedly change the way these scientists work and will help them find
new opportunities in the R&D of tomorrow.
We know that the journey from data to insights is long and complex. Rather than doing this all alone, we believe that partnership is the key. Elsevier has unique strengths to leverage: text-mining capabilities with our Scibite team (find here an example where they built a searchable database of breast cancer biomarkers); a data integration platform to deliver clean, normalized and contextualized data (find here aspects of data security, collaboration and high-quality data for chemistry modeling on the platform); and an entire team dedicated to address complex data challenges. We are looking forward to discussing with you any of the aforementioned topics and see how we can partner for the future.
References:
- From surviving to thriving: reimagining the post-COVID-19 return, https://www.mckinsey.com/featured-insights/future-of-work/from-surviving-to-thriving-reimagining-the-post-covid-19-return
- Reducing the Concepts of Data Science and Machine Learning to Tools for the Bench Chemist, Lewis et al., Chimia 73 (2019) 1001–1005
