Category Archives: metabolomics

Metabolomics integrates the effects of the environment with the effects of genetics

Metabolomics and Precision Medicine

Advancing Precision Medicine: Genomics, Metabolomics, and Clinical Trials

Monday, October 12 was the evening of an interesting talk at BIOCOM. Teresa Gallagher, founder of the San Diego Clinical Research Network (SDCRN) introduced the moderator of the event, Arnold Gelb, MD, Senior Medical Director at Halozyme. Rather than attempt to summarize all of the topics examined, the goal of this blog is to give a sampling of some of the areas discussed during the event.

Deterministic versus probabilistic genetics

The first speaker of the evening was Amalio Telenti, MD, PhD, Head of Genomics at Human Longevity, Inc. His talk touched on the ever-present nature vs. nurture debate. Do our genes determine a particular characteristic or merely influence the probability of developing that characteristic? In the world of whole genome sequencing, this can be described as deterministic versus probabilistic genetics.

In general, a deterministic trait would be something like Tay-Sachs Disease: if you have two copies of the gene for this condition, you have a better than 99% chance of developing the disease. A probabilistic trait is one with many genes that influence it, like height. Outside factors like disease and diet also affect how tall an individual grows. Hence, height is a probabilistic trait.

Telenti predicted that genomics will not revolutionize all aspects of medicine; but some medicine will be revolutionized profoundly; clinical trials will benefit the most. Genomics will be employed to stratify patient populations both before studies are commenced and after all the data is collected. Ideally genomics will be utilized to both determine who benefits from a drug and who should not take the drug.

Metabolomics combines genetics and environment

Steve Watkins, PhD, Chief Technology Officer of Metabolon spoke next.  Metabolon specializes in metabolomics, offering comprehensive measurements of small molecules such as glucose, cholesterol, cortisol, and amino acids in a CLIA-certified lab.

Metabolites reflect the integration of genetic and environmental influences on an individual.  Diseases can be prevented and diagnosed by checking on an individual’s metabolites. Response to disease treatment can be monitored by testing metabolites. Metabolomics is emerging as an effective tool in precision medicine.

Metabolomics integrates the effects of the environment with the effects of genetics

A person’s genome and environment affect their metabolome. Used with permission from Metabolon.

Watkins shared that Proceedings of the National Academy of Sciences recently published a study led by Baylor University’s Tom Caskey, MD. Caskey comprehensively tested the metbolites of many patients with no frank disease.  Metabolon’s platform spotted underlying health issues not previously noticed in the patients’ genetic data.

For example, Patient 3905 had very high levels of sorbitol and fructose, but no clinically significant mutation was reported in their genome.  Looking back at the genomic data for that individual, a mutation in the fructose pathway indicating “fructose intolerance” was discovered. This mutation had been overlooked previously. When discussing these results with the patient, the patient simply stated that fruit bothered him, so he refrained from eating it.

In the same study, Patient 3923 carried a gene for Xanthinuria type 1.  He showed no symptoms of the disease such as kidney stones, suggesting the gene was not penetrant (or not expressed), leaving the patient symptom-free.

In conclusion, Watkins stated that metabolomics can be used in a number of ways:

1)  By identifying pathways of interest for genetic assessment

2)  By revealing non-penetrance of genes suspected of being deleterious

3)  By enabling monitoring and understanding of metabolic conditions

Which drugs to use in cancer treatment?

The final speaker for the evening was Nicholas Schork, PhD Professor and Director of Human Biology at the J. Craig Venter Institute. He focused on emerging themes of design for precision medicine trials.

Schork presented several novel ideas. One was the idea of vetting algorithms for the treatment of cancers based on the mutations the cancers carry. Some hospitals already use this method, begging the question of who has the best algorithm for cancer treatment. As Schork points out, this has led to some interesting conversations with the FDA. He envisions clinical trials in the future for the evaluation of algorithms for cancer treatment with existing drugs, in direct contrast to the conventional clinical trial, usually designed to assess the effectiveness of a new drug.

In all, this was an exciting presentation of cutting-edge research and future directions in precision medicine.

Yes, these are lipids. But there are so many more inside your body; and they do more than store fat!

Annual Lipids Meeting in La Jolla California

The 2015 meeting on Lipids—focusing on their impact in cancer, metabolic, and inflammatory diseases—took place on Tuesday and Wednesday, May 12 and 13 at the Scripps Seaside Forum at UCSD’s Scripps Institute of Oceanography (SIO). With a beautiful venue and superb facilities, what more can you ask for? How about some really interesting science.

Lipids are generally thought of as fats. But in a biological system, they are much more. They include chemokines and other signaling molecules involved in signal transduction to and from the cell membrane. Metabolically, lipids also play an important role. Innovation in technology allow the study all the lipids in an organism (yeast, bacteria, or animal), leading to a new field of study: lipidomics. Once again, UCSD is on the cutting edge, with an established program and website in the field.

ocean, palm trees, La Jolla pennisula, green lawn with white chairs; breakfast view for lipids conference

View from the Scripps Seaside Forum at UCSD’s Scripps Institution of Oceanography.

Michael Snyder, the keynote speaker, has subjected himself to a battery of “omic” studies including his personal genome, exosome, microbiome, epigenome, proteome, metabolome, transcriptome, auto-antibody-ome, as well as cytokines. Data from these samples comprise the “Snyerdome”. All this was done in the interest of personalized medicine. These studies were done not only at one time point, but over a range of times, making it longitudinal.

Mike sees the data providing insights into how to managing healthcare in healthy individuals to predict risk, diagnose, monitor, and treat the patient, in this case, himself.

“He has also combined different state-of–the-art “omics” technologies to perform the first longitudinal detailed integrative personal omics profile (iPOP) of person and used this to assess disease risk and monitor disease states for personalized medicine” (from lab website).

in the future, Mike sees genomes being sequenced before birth and all this information being channeled through your smart phone. Patients will also bear more responsibility for maintaining their health with all the information they have; they will need to learn to maintain a balanced life.

The next speaker, David Wishart, discussed how to link lipidomics to laboratory medicine. He noted that in the rationalization of translating basic research to something of value in the clinic, researcher often cite the possibility of developing a new:          

  • surgical technique
  • invent a new medical device
  • drug
  • drug target
  • medically important gene
  • biomarker

All these are good outcomes; some are more likely than others. Practitioners of lipidomics are most likely to have the best luck in developing new biomarkers; not many are surgeon and drug development has about a 0.001% success rate from basic science to the prescription bottle.

lipids, lipids, lipids

Slide from David Wishart’s talk listing the number of FDA approved clinical tests from omic data

Discovery of new biomarkers is a realm where omics, specifically lipidomics, will meet a great chance of success. For this comparison, David recommends using the statistical ROC test, which is routinely used to evaluate medical test. This test gives a good sense of a medical test’s specificity and sensitivity by plotting the true positive rate over the false positive rate.

Example of ROC curve with an assessment of the area under the curve. The PSA referred to here is the amount of Prostate-Specific Antigen test; phi refers to a different, more specific calculation with less false positives than the PSA test alone.


ROC curve used to show predictive value of a test for prostate cancer using two different methods.

ROC curve used to show predictive value of a test for prostate cancer using two different methods.

Or you can just know that an ROC of 0.5 is worthless, while 1.0 is perfect.

Thus, from the graph above, using the PSA test alone to determine the risk of prostate cancer is poor. A better method is to use the phi method.

Work done looking at 3 to 5 biomarkers can have great ROC results. For example, predicting congenital heart defects by looking at the level of 3 carotenes, yields a ROC of 0.98. Other areas of success with high ROC scores include endometrial cancer, prostate cancer, and chronic fatigue syndrome.

David urged participants to become more quantitative to move their research into the clinic; using the website to generate ROC curves for your data is a great place to begin.

The numerous other speakers all gave fantastic talks.

In this smaller conference, I was able to browse through the all posters, read all the titles and talk to the presenters. Large conventions tend to lack the sense of intimacy and fraternity found in this lipidomics meeting. Kudos to the organizers for a successful event. A convivial group, I would highly recommend this meeting.