Leveraging AI for Drug Discovery in early stage Biotech Companies
Many biotech firms lack the basic data infrastructure to properly exploit digital tools and AI-technologies like machine learning, computer vision, entity normalization, network analysis, etc, and they don’t employ their own data scientists. However, some investors are specifically looking at biotech companies combining the DNA of both pharmaceutical and digital technology groups. Funding is only one reason why biotech companies will have to expand their skill sets to include the world of software, data, and service.
Since 2016, the biotechnology industry has faced unprecedented uncertainty in both Europe and the United States. That same year, biotech companies invested a record sum of capital into research and development, while revenue growth for publicly traded companies fell to 7%. New competitors like Apple or Alphabet might not be familiar with the regulatory hurdles, timelines, risks of therapeutics R&D, thus facing steep learning curves. However, they are quite advanced in their understanding of big data analytics and short cycle innovation; the skills biotech companies often lack and which are those necessary for sharing today’s life sciences landscape. At the same time as new players are entering the biotech industry, the ecosystem in Asia – China in particular – is maturing. Biotech companies will need to leverage and incorporate emerging digital technologies into their R&D processes, or be replaced by those who do.
When it comes to drug approvals, only one in five drug candidates make it from phase 1 to commercialization, while nearly 50% of compounds are lost in phase II and phase III trials for lack of efficacy. There are many reasons why drugs fail to be marketed – most common to both pharma and biotech companies – due to high competition by biosimilars or picking the wrong targets, to research failure, poor data quality, and not making the right decisions about compounds. As a result the industry must more aggressively address its R&D cost structure and improve development efficiency and effectiveness. The two aspects of science and strategy could be improved by better mining of the information and evidence that’s out there.
But how can conventional early stage biotech companies compete when commercial leaders have more capital?
A host of technologies, data, and analytics tools offer opportunities to address some of the ROI challenges by driving greater efficiency across the entire R&D value chain from early discovery to clinical trial management. This includes molecular data, positive as well as negative study findings related to compound efficacy, and abundant data on commercially relevant reimbursement and outcomes that can be geared to strategic decision-making.
Every day researchers navigate through hundreds of pages of publications, clinical trial outcomes, and congress papers. Sometimes the toughest challenge is not developing treatments but generating relevant and reliable real time insights from life sciences’ big data. Overcoming this challenge by way of AI will automatically and significantly reduce the time and money spent. Although AI will transform R&D productivity to some extent, radical transformation is to be seen regarding the efficiency at different phases of drug development, starting with streamlining components of drug discovery, allowing rapid screening of huge numbers of molecules for instance. Moreover, it will promote prospects of success at every stage of drug development.
Clinical trials are accounting for the largest expense in research and development, and new data-driven processes are available to enhance efficiency. Trial data is being digitized, and connected patients are accelerating trial recruitment; the identification of patients – likely to be responders to a particular drug – allows trials to be smaller, potentially reaching significance faster. First evaluations show that drugs developed with predictive biomarkers aimed at the selection of prospective responders are three times more likely to be approved than those without.
There is already an entirely new, growing biotech subsector that is built around the intelligent application and analysis of data. An emerging cluster of firms started to make use of artificial intelligence, powerful computers to identify links and patterns across vast quantities of data, to generate viable drug targets and leads – more rapidly than by conventional means.
Early stage biotech companies are, however, challenged by their own limited quantity of data, insufficient automated access to external data and their heterogeneous nature. Advanced analytics technologies are therefore required to produce useful insights.
Democratizing Life Sciences Data for Small Biotech Companies
Life sciences data is a treasure trove of insights, connections between biological entities such as genes, proteins, diseases and drugs, as well as information that can be leveraged 80% faster than manual approaches using artificial intelligence. Many small biotechs with limited knowledge and experience in data analytics, and moreover, much less data to analyze, are hardly able to generate useful or relevant information to foster their business decisions. Larger companies hold a data monopoly while smaller more innovative ones are struggling to catch the right information for the right context.
The amount of data available to these biotechs is so small that its analysis comes close to a ‘nothing ventured – nothing gained’ situation. It takes the whole picture to analyze the outcome! But there is a lot of valuable data especially for clinical trial and KOL management that is external and publicly available. However, a lot of third parties offer this data too, and it is expensive. Not to mention the expenses for an expert in-house data analytics team that understands emerging digital technologies like AI and machine learning to permit generating fresh insights and identifying new patterns from data.
AI-enabled tools on the other hand are capable of sifting through myriads of data to increase the efficiency of research, and give scientists relevant real time insights from fully automated search on life sciences data. These tools enable the smart discovery of information through interactive research graphs using big data of life sciences that would otherwise be hard to analyze and take up tremendous time.
Biotechs could leverage AI to predict the structures of molecules, accelerate research efforts, increase the potential of clinical trials, and to analyze relationships between various entities, generating interactive research graphs and creative dashboards with centralized information on, for example, KOL discovery, sentiment analysis, guidelines, etc. to stimulate drug discovery and faster development.
If scientists open themselves to the benefit of machine learning and data crunching in their fields of endeavor, and ask new kinds of questions previously intractable, the breakthroughs in the biotech industry will endure. Data scientists need to talk to researchers for a better understanding of how their tools are going to be used. If that happens, technology platforms like gene editing, cell therapy, and next generation sequencing will continue to emerge.
Boosting R&D efficiency by embracing emerging technologies, incorporating digital and artificial intelligence will be indispensable for biotech companies to simultaneously raise the return on investment and the affordability of drugs.