ARCHIVE. In 1997, a movie called Gattaca hit the screens and got everyone talking about what life would be like if you could predict a person’s future based on their genes. The ethical implications of precise genetic prediction would be massive, as babies would be judged harshly before birth. Research has come a long way since the 90s, however, and Gattaca is now more comic than compelling. Genetic differences, although precisely linkable to certain diseases, are limited in their ability to determine whether and when a disease will actually occur. The environmental and lifestyle variables are just too considerable and are where many researchers are focusing their efforts today.
“You know genes play a role in diabetes but you are not sure where along the line they do and how they interact with other factors, such as what you eat and whether you do a lot of sports,” said Karsten Suhre, Ph.D., Director of the Bioinformatics Core at Weill Cornell Medical College in Qatar. “There’s a big question in there, and that’s where my work comes in.
“We are working at the metabolic level,” Dr. Suhre said. “Metabolites are sugars, amino acids, everything in your body that the enzymes affect. So if you have genetic variation that determines a different enzymatic setup, of course that would lead to a change in the person’s metabolic makeup.”
This makeup, based on how substances are processed in the body, is known as a metabotype and is determined by the different metabolite levels at the time a sample is taken. This kind of reading has the potential to predict disease risk more accurately as links are made between blood levels and genetics. Dr. Suhre said there are presently 500 metabolites that can be tested against genetic variables. Glucose is only one of them and is a well-studied biomarker for diabetes.
Biocrates example of clearer picture
Evolution of disease resolution, courtesy Biocrates (click image to enlarge)
The study of metabolomics dates back to the middle ages, when urine samples would be tested against charts for their smell, color and taste. The premise remains the same today but the methods have become extremely sophisticated.
“It’s like taking a photo of a person, but from the inside, when you measure the metabolic state of an individual in a high-throughput manner,” Dr. Suhre said. “It hasn’t been done to date, because we haven’t had the technology in place to process thousands of samples and detect hundreds of substances in them. Now we do.”
“We have created here in Qatar what we call a virtual metabolomics lab. It is the same setup I had in Munich [working on the KORA, Cooperative Health Research in the Region of Augsburg, study]. However, I don’t have a real lab with instrumentation. Instead, I work linking the findings of two research groups together—epidemiologists [specializing in population-studies] and metabolomics lab people. I’m a bioinformatician focused on population studies, genetics and metabolomics. I connect these groups and do the computation and the data analysis.”
The study of humans is complicated and laden with challenges, with hundreds of factors coming together and only a limited amount of individuals sampled to date. In order to get higher-resolution pictures of how genetics and disease relate through metabolic processes, the sample sizes studied must be very high, Dr. Suhre said.
“People are not lab mice,” he explained. “In lab mice you can knock out one gene and see the signal but in humans you obviously can’t do that. There is a lot of individual variation. That’s why as hundreds of factors come together it’s hard to find a clear signal that can be related to only one of them.”
Dr. Suhre said there are two main approaches to getting a clearer picture about metabolic relationships to disease and genes: “A genome-wide association study of the relationship between SNPs [Single Nucleotide Polymorphisms: isolated genetic differences] and metabolomics, and a metabolome-wide association study of non-genetic factors, such as coffee consumption, nutrition habits, smoking, medication, physical activity and other lifestyle choices and how these link to a person’s metabotype.”
Last fall, Dr. Suhre and colleagues were published in the journal Nature (43:565-569, 2011) for their findings related to 37 SNPs that were strongly linked to metabolic traits, 23 of which had never been seen before. The study was based on blood from 2,800 people.
“Two things are critical in this field of study, one is you increase the sample size," Dr. Suhre said. “The other thing is to increase the intermediate metabotypes studied in relation to the SNPs, because even with large sample sizes, the genes don’t show enough. I can measure a million genotypes at a time, but many of those have no effect. If I roll seven dice a million times, I’m bound to reach a point where I eventually get all sixes, although such a pattern is expected.
“Likewise, at some point, if I have only a small number of study participants, I may find that all of my diabetics have one genotype and the healthy ones have another, but this could be a chance event,” he continued. “If I increase the number of individuals in my study, at some point it cannot be a chance event anymore; let’s say instead of seven we have a hundred dice and roll them a million times, then rolling an all-six pattern would be extremely unlikely.’”
Dr. Suhre is currently collaborating with many researchers at WCMC-Q and Qatar's national health care provider, Hamad Medical Corporation. Many of these collaborations are supported by QNRF National Priorities Research Program grants. Genetic analysis takes place mainly at WCMC-Q, and and some of the analysis of the metabolite data and lifestyle and symptom information is being coordinated with Qatar Computing Research Institute and Texas A&M University in Qatar.
“Genes remain the same over your lifetime, but metabolites change all the time,” Dr. Suhre said. “Study parameters, the day a sample was taken or whether or not a person was fasting, all kinds of factors affect metabolites and must be considered, controlled and documented. So we are coordinating with our clinical research partners to ensure that the samples are standardized so that the data we work with into the future is of the highest quality we can achieve.”