SAN DIEGO – Data is the currency of the industry and masses of it are generated every day adding to the rapidly swelling repositories of public accessible and private databases around the world. The problem, however, remains that biopharma companies are whizzes at generating big data but this hasn't, as yet, helped them improve upon the number of new medicines that they bring over the goal line annually. There is a growing recognition that if we are able to successfully interrogate the disparate treasure trove of information with sophisticated algorithms the resulting output could assist companies in drug development and ultimately reduce the burgeoning costs of this activity. This is why companies are beginning to pay a great deal more attention to machine learning and artificial intelligence (AI), techniques which are themselves also advancing to reach a point where these computational tools have the ability to contribute to the improvement of drug development R&D.

The subject of AI and its role in biotechnology was a prominent theme in many panels that were on tap during last week's BIO 2017 International Convention.

Not wizardry

Demystifying the notion that AI is the stuff of gurus or wizards was Atul Butte, director of the University of California, San Francisco, Institute for Computational Health Sciences, and a participant on two of the panels devoted to the potential of AI in drug development. It is merely sophisticated software that, in combination with machine learning, can be "trained" to find patterns or uncover "knowledge gaps" among masses of data, he explained.

Butte is passionate about the potential of data and its role in drug research. AI can completely alter the way we think about drug research. For example, we can begin a clinical project with existing data on patients in the area we are interested in rather than going out and getting samples from patients from the get go.

Garage biotech

Butte introduced the concept of garage biotech, where anyone can start a company using the existing open data repositories. He cited several examples of this including Carmenta Inc. and Numedii Inc., which have enjoyed commercial success in a relatively short period by analyzing publicly available big data to develop products.

Menlo Park, Calif.-based Numedii is focused on the discovery of effective new drugs by translating predictive data intelligence technology into therapies. It does this by applying its biological network-based algorithms on scientific data. The technology employs deep learning of human biology consisting of hundreds of millions of unstructured public, molecular, pharmacological and clinical data points that the company said it has curated and harmonized. The technology is designed to uncover drug-disease connections and biomarkers that are predictive of efficacy. The predictions become the starting point for the development of de-risked drug candidates.

In May the company formed a strategic partnership with Three Lakes Partners LLC, to discover and advance new treatments for idiopathic pulmonary fibrosis (IPF) based on their intelligence technology.

The company has established discovery collaborations with three biopharmaceutical firms including Astellas Pharma Inc. where it is using its technology to identify new indications for a number of undisclosed Astellas compounds.

Testing the waters

It's only recently that AI has started to make inroads into health care and some large biopharmaceutical companies have just started to test the waters, Glen Giovannetti, EY Global Biotechnology Leader told BioWorld Insight.

His firm released its 31st annual edition of Beyond Borders at BIO 2017, which, among other issues, found that the industry continues to be challenged by R&D productivity.

In addition industry is shifting "from a clinical science supported by data to a data-driven science supported by clinicians," he observed. The shift requires companies to adopt emerging technologies such as AI to remain productive and relevant.

"Artificial intelligence and the accompanying analytics are now so advanced that these tools promise to improve the traditional drug target selection and R&D process. However, how biotechs, especially smaller ones, optimally access these capabilities remains an important question," the report says.

One company that has started along the AI path is Pfizer Inc. It has signed a collaboration with IBM Watson Health to put its research in immuno-oncology on the fast track. IBM said that the big pharma is one of the first organizations worldwide to deploy Watson for Drug Discovery, and the first to customize the cloud-based cognitive tool to tap in to Watson's machine learning, natural language processing, and other cognitive reasoning technologies to support the identification of new drug targets, combination therapies for study, and patient selection strategies in immuno-oncology.

A participant on the panel entitled "AI – Opportunities and Challenges in Transforming the Biopharma Value Chain," Iya Khalil, chief commercial officer and co-founder of Cambridge, Mass.-based GNS Healthcare Inc., said that AI is all about combining established mathematics with the power of supercomputers and making sense of often complex, messy data.

Her company disclosed a collaboration with Basel, Switzerland-based Roche AG's Genentech unit to leverage GNS REFS (Reverse Engineering and Forward Simulation) causal machine learning and simulation platform to power the development of novel cancer therapies. The GNS technology turns large and diverse patient data streams into mechanistic computer models that reveal new pathways, novel targets and diagnostic markers that may lead to new treatments that are better matched to individual patients, the company said. The duo will join forces to unlock knowledge from various data resources such as longitudinal electronic medical records, next-generation sequencing, and other data.

Training your computer

Speaking at the Digital Health Summer Summit, which was co-located with BIO 2017, Peter Neubeck, chief medical officer at ExB GmbH (standing for External Brain), described the company's technology platforms in non-linear programming and deep learning. These are being applied to create machine learning solutions for a range of health care problems through its subsidiary ExB Health. The founders of ExB Labs were one of the first to develop and implement text-prediction software for cellular phones, which they eventually sold to Nokia.

Neubeck said that through its cloud-based platform, the Cognitive Workbench, users are able to create and train their own AI-enabled analyses of complex unstructured and structured data sources with "human precision."

The platform comes with a range of publicly available data sources such as PubMed, patent databases, etc., which can be combined with the users' data, which is often stored in unstructured format such as research reports, health records or images. Using a number of solutions this data can be made available for analysis.

The Cognitive Workbench has been designed to recognize patterns in research and medical data sets appearing in both text and graphic forms. Customers can customize their search criteria leading to better pattern recognition, hypothesis generation and signal-to-noise ratio. Over time, with repeated inputs, a more precise output is eventually obtained.

Last week, the company reported that it had entered into a strategic collaboration with Alacris Theranostics GmbH, a systems medicine company, focused on exploiting next generation sequencing and other omics data through its predictive modelling system Modcell, to support drug development and personalized medicine in oncology and other complex diseases. The technology was originally developed at the Max Planck Institute for Molecular Genetics in Berlin and is exclusively licensed by Alacris.

The goal of the collaboration is to offer customers a means to accelerate the evaluation of their omics and drug-response data, for the identification of responders and for stratification of patient or experimental model cohorts.

Going forward

There is no doubt that we will see more biopharmaceutical companies partner with AI firms. There is a strong imperative for them to improve their R&D efficiencies because as the EY Beyond borders report points out, "unless pharma can start to reduce R&D costs – and time – ROI will eventually fall to levels that threaten the sector's viability."