A genome is essentially a database of a living thing — which makes translating this precise information a powerful approach for guiding drug development and bringing treatments to market faster.
For instance, consider the role of genomics in clinical trials in the emerging field of immuno-oncology.
“Many of the trials in immuno-oncology include a genomically targeted therapy,” says Anita Nelsen, Head of PAREXEL® Genomic Medicine Services.
With advanced analytics and predictive data modeling, Nelsen says, “researchers are changing the design of clinical trials to enable patients to be selected based on the genome of their tumor and to help streamline patient recruiting and decision-making during the clinical trial.”
As we move toward more targeted treatments for cancer there are key questions to consider. For example, what’s the mutational burden across the tumor and how is it affected by a drug treatment? As the tumor grows — or if it metastasizes — how do the genetics differ across the heterogeneous tissue or across tissues within the same person being treated?
By answering these questions, researchers can better understand the genomics of a tumor as it relates to the drug target and drug response. And that’s just one scenario out of the millions of possibilities raised by genomic medicine and data science.
As researchers continue to pinpoint the relationship between a drug, target, patient and disease, they can further help to inform drug development in cancer and across a wide range of conditions. Biopharma leaders confronting these possibilities should keep three key points in mind:
1. You Must Master Two Distinct Disciplines
Data scientists pull together large data from multiple sources, including genomic studies and create algorithms that enable models that compare the impacts of multiple scenarios. Subtle refinements in the model can make it more predictive.
Genomic scientists, by contrast, may have expert knowledge in the biology but may not be well-versed in the data and caveats that are associated with leveraging big data. Without the right expertise, you may find a gap that can bedevil attempts to merge data science and genomics in clinical trials.
“Those involved in developing the algorithms behind the technology may often lack genomic domain knowledge. In converse, many genomic scientists may not be well-versed in the data and the caveats that are associated with leveraging big data,” Nelsen says. “Who can cross this bridge? Those who have an understanding of both the technology and the science.”
2. It’s Not Just About the Data — It’s About the Metadata too
Metadata is key to making sense of gigantic genetic datasets — enabling you to understand the data and therefore to understanding its quality.
Genomic technologies have advanced rapidly over the last decade and we have made significant strides in being able to computationally interrogate this data to better understand the biology of disease, drug targets and response to medicines. During the course of discovering and developing a drug, data scientists may analyze data generated across multiple analytes, platforms and populations.
“Understanding the way in which the data was generated and what the individual data values represent is critical,” Nelsen says. “Think of metadata as the data about the data. Metadata gives you the information you need to understand the context from which that data was generated.”
Metadata can help you make sense of factors such as patient consent, specimen quality and handling, technologies used in the laboratory and even how the data is formatted when it is collected.
It’s crucial to have a solid metadata foundation given that data modeling will compare, contrast and attempt to predict based on the signals provided by your data.
3. Genomics May Help Reduce Attrition and Streamline Development
According to a recent article in Nature Reviews Drug Genetics1, approximately half of phase III trials fail due to lack of efficacy. In contrast, drugs with genomic evidence linking the drug target and disease indication are twice as likely to succeed to approval.
By analyzing genomic and other biomarker data in the public domain, as well as exploratory biomarkers collected during clinical development, researchers aim to gain a better understanding of the drug target, the disease(s) and populations most likely to benefit from treatment with the drug. Biomarkers or biosignatures could then be used to identify populations in future trials, thereby streamlining recruitment and reducing the number of study subjects needed to demonstrate efficacy.
“By identifying patient populations most likely to benefit from the drug, we may be able to streamline recruitment and get your drug to the right patients faster,” Nelsen says.
Genomics and statistical modeling can make an impact throughout drug development, including during regulatory approval. Nelsen points out that, more frequently, regulators are requesting genomic evidence to help them make decisions.
“Companies should certainly consider that there is still opportunity for genomic influence on a trial, even if you're not starting with a genomic strategy when you first commit to your human studies,” Nelsen says.
1 Nature Reviews Genetics, The genetics of drug efficacy: opportunities and challenges, published online March 14, 2016