Sorry, you need to enable JavaScript to visit this website.

Artificial Intelligence in Life Sciences: Overcoming Roadblocks along the way

Artificial Intelligence in Life Sciences: Overcoming Roadblocks along the way
August 16, 2018

Most industries, including life sciences, are witnessing a transformation owing to increasing cost pressure, a greater need for productivity, and disruption caused by new and innovative market players.

As life sciences organizations look forward to doing cutting-edge research, discover new drugs faster, or stay ahead in the race of latest equipment, technology has emerged as the single biggest enabler in tweaking their value chains to meet their current business requirements.

One such technology enabling this industry is artificial intelligence (AI). AI is a branch of computer science that emphasizes the creation of intelligent machines that work and react like humans. As of now, it has a little, yet developing impression in this industry with increasing application in drug discovery and development. But firms are expanding AI applications across the product lifecycle. This industry involves a huge amount of data, structured and unstructured, and to manage this data better AI is the recourse. It also helps improve the processes and save time and cost.

Life Sciences industry involves a huge amount of data and to manage this data better AI is the recourse which also helps in improving the processes and save time and cost.

But applications of artificial intelligence do not come without their set of challenges. Integration of AI talent and technology is one of the major concerns pharma companies face, though talent is the most significant challenge. When compared to experts in other fields, there are a handful of people who have expertise in AI, ~10,000 [Shelley Zhuang, Founder and Managing Partner at 11.2 capital in an interview to Techemergence] in number and even fewer with biopharma or healthcare experience. To solve meaningful problems with the applications of artificial intelligence, interdisciplinary innovation and collaboration is a must. The other challenges include but are not limited to:

Adoption Mindset

Despite huge R&D budgets, the life sciences industry has always been slow in adopting cutting-edge technologies. Since this concerns patients’ health and safety, multiple checks are required at different levels involving huge data sets that impede timely deployment of the new technology. They need to validate the intended purpose of the technology before applying it in their processes.

“According to a survey of 229 life science professionals conducted by The Pistoia Alliance, around 72% of them believe their sector is lagging behind other industries in its development of AI.” – Pistoia Alliance

Data Complexion

Pharma companies sit on a huge pile of data. Due to the competitive operating environment and strict regulatory compliance standards, this data is hidden behind a firewall and is sealed from the rest of the world. This makes it tough to access and work with, for applications of artificial intelligence. Also, it is highly confidential as it contains the patient’s health records, involving their medical history, personal information, and other crucial details. So, sharing this data with partner firms becomes difficult and is not considered suitable for retrospective analysis.

Talent and Capabilities

The thought process of a biology expert and a computational expert is starkly different. Also, computational biologists with wet lab experience are rare to find. One needs to understand the biological processes when working in the life sciences industry, and currently the AI pool of resources lacks that expertise and/or knowledge. The need of the hour is a balanced approach to understanding biological experiments (how a certain result is achieved), along with a hacker mentality which celebrates agility (speed over understanding underlying process).

Even though there are quite a few significant obstacles to integrating AI in a life sciences environment, it has not put brakes on the efforts of the pharma companies.

Pharmaceutical firms look at solutions which can be mainly divided into external and internal.

Externally, pharma companies are leveraging R&D capabilities of innovative tech companies equipped with mature and stable data, and following a technology-driven approach, by either partnering with them or by acquiring them at an early stage of their evolution. Pharma companies are even eyeing ready-made solutions designed by such firms. Another model that is emerging is the partnership between pharma companies and med techs (medical technology firms including medical device manufacturers).

“Given the talent shortage reality, not every life sciences organization will be able to justify the cost or have the time to build an AI practice in-house. One way to bridge this gap is to partner with other companies” – Shelley Zhuang, Founder and Managing Partner at 11.2 Capital

To enable full-scale digital transformations, pharma companies are looking internally for IT infra. They are devising different strategies to enable this transformation. For example, as a norm, leading organizations are turning to low-code application development platforms. It not only speeds up the delivery of applications but also maintains transparency, security, and traceability throughout. Also, for curated data, they are entering into more robust data contracts with the vendors. This enables them to incorporate advanced analytical programs to extract insights effectively, required for strategic decision-making.

Pharma companies are already working on a few promising ideas to overcome these roadblocks:

  • Forming separate “data science” teams apart from their conventional IT efforts. They are trying to include experts from the fields of bioinformatics, clinical analytics, machine learning, etc. The main focus of these teams is to integrate computer science, AI, and life science.
  • Creating platforms to provide easy access to structured data to all parts of the organization and for varied use cases.
  • “Creative destruction” of conventional barriers between different fields such as research and development to enable cross-functional and divisional teams, which helps them overcome traditional limitations and knowledge barriers.

Pharma companies are being careful by not doing all the innovation themselves and are opening up to technology firms and start-ups and outsourcing a significant part of the pipeline. For example, LEAPS from Bayer. Also, they are establishing new R&D collaborations to counter the talent shortage, such as AstraZeneca with Berg Health.

However, there are a few pointers that pharma companies should bear in mind:

  • Data from a single source or single type of data (e.g., genomics) does not provide the best insights. Rather, an amalgamation of various data types and the insights generated from them at the same time is required (e.g., multiomics versus just genomics)
  • To shift the focus to technology and its use cases vigilant efforts in recruitment and retention of the AI talent is necessary
  • While the influx of AI in other fields is fast-paced and perverse, the life sciences industry is yet to embrace AI applications with open arms. Being a heavily regulated industry, an open mindset is critical for the adoption of AI in healthcare and to overcome other challenges, such as the dearth of talent and data complexity.