How computational biology is transforming modern drug discovery

In recent years, computational biology has been going through something of a renaissance. Despite being a field that is several decades old, the information-rich resources that are available today, in combination with modern computing powers, are enabling computational biologists to delve deeper into and explore the broader aspects of some of the most complex biological questions. It’s certainly an exciting time in the field.

To provide those in the industry with unique insights into the future of computational biology, we asked Professor Bissan Al-Lazikani, Chair of Cancer and Drug Discovery Data Science and Head of the Department of Data Science at the Institute of Cancer Research, to present at our upcoming Drug Discovery 2018 conference (October 9th and 10th at the Excel Arena in London – (you can reserve your free place at the event here).

In an exclusive preview before the event, Professor Al-Lazikani spoke to us about the current state of computational biology, how the techniques used can be employed in drug discovery research, and how the approach could ultimately serve as a weapon in the fight against diseases like cancer.

How computational biology is rising to the challenge of abundant data

Today, the amount of data we produce is increasing exponentially and biological information is no exception. In fact, sources of biological data are becoming so abundant that there can often be an overload of information for researchers and healthcare professionals, and drawing actionable conclusions is becoming more difficult. As such, computational biology has emerged as a key way to find patterns in these vast datasets in order to unlock the insights within.

As Professor Al-Lazikani explains: “In the biological sciences, a technological revolution has taken place over the past twenty years. Advances such as the human genome project, accurate modelling of the human brain, and developments in molecular profiling technologies have brought with them a wealth of detailed data, and as such, a fresh avenue of exploration for computational biology.” However, with this unprecedented increase in the volume of data, we also require the application of sophisticated mathematics for analysis.

An important area where computational biology and data science have risen to prominence recently is in drug discovery. This is because effective drug discovery requires a complex understanding of many disciplines – from the underlying chemistry and the diverse biological makeup of the patient population, to the biological pathways involved and the genome underpinning them. So, by finding connections between the distinct specialisms and domains of science, computational biology methods can help drug discovery scientists make more accurate decisions.

“Multi-disciplinary analysis using all the available datasets can really help drug discovery scientists across every step – from the generation of hypotheses, to the identification of novel targets, to drug design,” says Professor Al-Lazikani. “But computational biology can go even further by helping you plan for a Phase I clinical trial, including identifying biomarkers for patient selection. All of this is geared towards giving your drug the best chance of success.”

How can computational biology help overcome the challenges facing modern drug discovery?

“Drug discovery is a balancing act between risk management and innovation,” Professor Al-Lazikani says. “In an attempt to mitigate risk, drug discovery research often focuses on well-established targets and then refines around them. However, it’s the novel, riskier targets that will help us cure diseases in the future.”

Computational biology can define the risks early on in a discovery programme and help researchers select target candidates that have the lowest potential for failure. In this way, they would be able to innovate with more confidence. This approach allows testing of risk and abandoning programmes with little chance of success earlier, and before the hefty time and financial investment.

As Professor Al-Lazikani states: “All of this feeds innovation. I believe that the full application and utilisation of computational technology will speed up drug discovery, and will ultimately give us better, more innovative drugs in the future.”

Have data analysis and AI already had a positive impact on drug discovery?

“Absolutely!” Professor Al-Lazikani says. “For instance, at the Institute of Cancer Research, we have developed a public resource called ‘canSAR’ to help with drug discovery. We have been able to demonstrate that by applying sophisticated multi-parametric studies, you can uncover areas for drug discovery that would have been completely neglected otherwise.

“As an example, we have published an application of this approach where we employed a suite of machine-learning algorithms to genomic data to select novel drug targets. We identified 46 targets that looked druggable but that hadn’t been used for any previous drug discovery research, or even any serious medicinal chemistry investigations. I’m happy to say that over 10 of these targets are now the subject of active drug discovery programmes.”

The importance of open access data resources and precompetitive data sharing in drug research

A key point to note is that, if computational biology is to be used for effective drug discovery, it requires access to as many relevant resources as possible. As such, the sharing of knowledge, data, and expertise is crucial to fuel progress in the field and will accelerate innovation accordingly. Fortunately, there is a growing appetite worldwide for ‘precompetitive’ data sharing, with resources like canSAR an illustrative example of what can be achieved.

However, it’s not only the Institute of Cancer Research that sees the benefit of open access resources. As Professor Al-Lazikani explains: “We are working with organisations like the Broad Institute and the Sanger Institute who have fantastic expertise in functional profiling. We collaborate with them and are working on ways to not only share data between each other, but also to combine our data into a single engine. This is so that the rest of the world can benefit from our integrated capabilities.

“You might ask why we are giving away all this knowledge about drug discovery when we could be using it ourselves. Well, the answer to that is there’s really no alternative. The problems we are trying to tackle are so big that they require the sharing of knowledge and expertise precompetitively and openly. Fortunately, I’m really happy to say that most of us who work in this field are really buying into this idea of precompetitive data sharing.”

Computational biology is dramatically benefitting oncology therapeutics

The treatment of cancer, perhaps even more than other diseases, is benefiting from this modern explosion of data. It is possible to collect genomic, gene expression, protein expression and histopathology data about patients in numbers that would have been unimaginable previously. As Professor Al-Lazikani says: “This brings with it the opportunity to really understand the disease and find new therapies for it.”

In a recent Channel 4 interview, Dr Al-Lazikani expresses her confidence that AI and machine learning will enable more timely diagnosis of cancer and the discovery of new treatments. “Even those types of cancer that are the most challenging to cure will become more manageable through the smart adaptation of therapies to individual patients,” she adds.

This is why Professor Al-Lazikani thinks this is the most exciting time to be at this interface between data science and cancer research. “I can’t imagine a better time or a better context in which data science can really transform the lives of literally hundreds of millions of people all around the world – simply by applying these innovative data science technologies to the vast datasets that we already have access to,” she says. “I certainly wouldn’t want to be anywhere else.”

Computational biology, emerging technologies and the future of drug discovery

Currently, the validation of a target can be a real hurdle in the early stages of drug discovery. As such, Professor Al-Lazikani says that something she’d really love to see is “the utilisation of computational technology combined with robotics and other experimental technology advances to overcome this important challenge.”

To this end, Professor Al-Lazikani envisions a future where “artificial intelligence methodologies are able to immediately devise the precise experiment for each particular target and inform the experimental lab. Then the AI could combine with robotics and better lab technologies to enable the validation of targets at a much faster rate than we are able to do now.” This would ultimately accelerate progress within the field. It also paints a picture that suggests a very exciting future for everyone involved in the drug discovery sector.

Book your free place at ELRIG Drug Discovery 2018

Computational biology is just one of the topics we’ll be covering at this year’s Drug Discovery 2018 meeting (October 9th and 10th at the Excel Arena in London).

We’d obviously love for you to be there to share it with us, but we think Dr Al-Lazikani sums it up best when she says: “Programmes like Drug Discovery 2018 are really exciting because they open up the door for bringing in new ideas which can then massively impact drug discovery. I think the event will allow people to really witness how fluid and how different drug discovery is now from what it was just a decade ago, as well as what an exciting field it continues to be.”

To learn more about the future of drug discovery, book your free place at ELRIG Drug Discovery 2018 now.