Jeffrey S. Buguliskis Ph.D. Technical Editor Genetic Engineering & Biotechnology News

Tackling the Glut of Data Will Be the Key Challenge

It seems every week that a new publication extols the benefits of next-generation sequencing (NGS), often from sequencing the genome of a medically relevant organism in record time or identifying the polymorphisms that may lead to drug resistance in cancer. One could make an educated guess that NGS is poised to tackle almost any clinical issue facing precision medicine. But what about the glut of genomic data that is unavoidably generated from patient samples with each sequencing event? Is there relevant data from those samples that is clinically actionable, disparate from the initial clinical presentation? These are some of the current questions that physicians and clinical researchers will have to face as they foray into the personal genomics era.

With the ever increasing availability of direct-to-consumer genetic testing, coupled with the accelerated fall in costs for whole-genome and exome sequencing (WGS and WES), the opportunity for individuals to analyze their own genomic data and take greater control of their healthcare decisions will be abundant. Patients will want to know what genomic biomarkers are and how they are relevant to everyday health or their current medical condition. Physicians will need to guide patients through the genomic data maze and explain why some types of information might be clinically actionable and others inconsequential.   


To Define Is to Limit

A biomarker is broadly defined as an objective indicator that can be measured accurately and reproducibly for a medical condition from outside a patient, usually with little to no invasiveness. Biomarkers are used diagnostically since, in contrast, medical symptoms often tend to be subjective signs of health or illness as perceived by patients themselves. In the genomic age, the identification of clinical biomarkers has exploded, as they are not only used for diagnostic purposes, but many have become druggable targets for therapeutic intervention.

Similarly, a genomic biomarker is a measurable DNA or RNA characteristic that is an indicator of either a normal or pathological state. Although they are often employed as a verification tool in clinical diagnostics, genomic biomarkers are being utilized more frequently, once validated, in the early detection of disease states.

Two key concerns that many clinical researchers currently have are how to determine valid targets and when to act upon them. There are no current universally accepted recommendations that direct physicians toward a clinical plan should incidental findings arise from genomic testing. However, the American College of Medical Genetics and Genomics recently released guidelines regarding the reporting of incidental findings within sequencing data and the FDA does have a biomarker qualification program, but neither of these panels are comprehensive nor fully adopted by all institutions.

“The definition of actionable data is a key problem,” explained Jason Park, M.D., Ph.D., assistant professor in the department of pathology at UT Southwestern Medical Center. “Until the rise of genomic testing, actionable genetic testing was restricted to the tumor of origin. For example, HER2 amplification in breast or gastric cancer is a finding that is targetable with anti-HER2 agents such as herceptin. However, finding HER2 amplification in other tumor types was not considered a valid target.”  


Participation May Vary

Clinical institutions currently find themselves in a very difficult situation. On the one hand, most would like to take position at the frontiers of precision medicine, touting their expertise in both personnel and the latest equipment. On the other hand, many institutions have only just begun to tackle the issues surrounding the clinical actionability of incidental findings from WGS or WES studies.

“Our institution has a Center for Individualized Medicine and one of the areas of emphasis is biomarker discovery,” states David Smith, Ph.D., professor of laboratory medicine and pathology at the Mayo Clinic. “Unfortunately, our institution does not have a plan in place to determine which biomarkers are therapeutically relevant and worth pursuing in patients. Part of the problem is a disconnect between discovery and clinical translation.”   

Similarly, when asked about the current state of affairs at UT Southwestern, Dr. Park mentioned that they have only just recently started a genomics tumor board to address the actionable biomarker issue. “At this monthly meeting we discuss the results of genomic or multigene tumor tests and implications for therapy and/or genetic counseling.”

Additionally, Dr. Park wanted to leverage his own research in an attempt to address this issue. “My clinical laboratory provides a 25-gene-targeted oncology-sequencing test that examines genes with therapeutic and/or prognostic implications. For both the genomics tumor board as well as the in-house test, the therapeutic choices are limited to what has already been cleared by the FDA or what is currently part of a clinical trial.” 

Even a cursory scan of current clinical publications would lead anyone to the conclusion that the Mayo Clinic and UT Southwestern Medical Center are far from the only institutions facing difficulties in resolving this complex issue.


Could Incidental Findings Derail the System?

A common scenario exists in clinical medicine—a patient comes to the hospital complaining of pain. An X-ray or CT scan shows a cracked rib, as well as an additional abnormal finding, a diffuse mass without clinical symptoms. This type of coincidental finding, often dubbed incidentaloma, is becoming more universal as diagnostic imaging techniques, such as whole-body CT scans are used more frequently.

The conundrum is what steps to take next. For many the answer would seem obvious: treat the mass. However, physicians are trained to order tests carefully and only do so if results would alter the management of treatment. Furthermore, data is beginning to mount suggesting that many of these incidentalomas are normal within the lifespan of humans. For example, 7% of patients over 60 may harbor a benign growth on their adrenal glands and up to 37% of individuals that receive a whole-body scan may have an abnormal discovery.

So common are incidental findings within diagnostic medicine that the phenomena has spread into genomic analysis techniques, significantly entrenching itself and threatening to undermine the progress of personalized medicine. The “incidentalome,” as it is often referred to, is a set of genomic biomarker variants that surface within the results from WGS or WES that were not originally part of the study design and it derives its name from the previously mentioned accidental findings of diagnostic imaging.      


Data Analysis Controls the Future

Ultimately it will fall on clinical researchers and physicians to make responsible decisions concerning the validity of clinically actionable targets. Interestingly, the bottleneck in the validation process rests almost exclusively with data analysis, which currently seems to be a common thread of many NGS endeavors. Yet many researchers are still enthusiastic about the information these genomic techniques have the capacity to unlock for various debilitating diseases. 

“While there will be a considerable log-jam when all this data is generated,” Dr. Smith stated. “I am optimistic that having this type of data is going to provide considerable opportunities to develop better ways of analyzing this information.”

It seems almost unfathomable in this age of smart phones and cloud computing that computer-aided data analysis would be, in large part, contributing to the congestion within the NGS data stream. Although, with genome assembly data files reaching into the hundreds of terabytes (1 terabyte=1,000 gigabytes=1,050,000 megabytes), it’s not difficult to see why NGS data analysis has been met with much consternation by sequence facilities. Nonetheless, if history is any indicator of the future, the computing side of data analysis will not remain the limiting factor for very long.    

“In terms of discovery, there certainly is a data issue, but I am confident that computing power will continue to meet the data needs of genomics,” concluded Dr. Park. “The growth in useful data will be capped by the growth in useful targeted therapeutics.” 






































Jeffrey S. Buguliskis, Ph.D. (jbuguliskis@genengnews.com), is Technical Editor of GEN and Clinical OMICs.

This article was originally published in the April 2015 issue of Clinical OMICs. For more content like this and details on how to get a free subscription to this digital publication, go to www.clinicalomics.com.

This site uses Akismet to reduce spam. Learn how your comment data is processed.