Interpreting Genomic Art

Interpreting Genomic Art

What the …?? …. Standing back, I wonder what the artist envisioned while painting this piece. Did they just enjoy seeing the colors mix and build elaborate textures? Are they trying to communicate a new revelation about the human condition?

It turns out the artist is barely two years old. Far too young to really be grappling with existential questions. The artist most likely just enjoyed playing with the paint.

We have a normal tendency to over- or misinterpret things, whether it be art itself or the artist’s intentions. We can do the same with genomic data. With over 75,000 genetic tests available (Johnson, 2018), the deluge of information can wash over us and the medical experts, leaving us grappling with deeper questions of meaning and use.

Recently, a study came out about how primary care doctors, cardiologists or oncologists view genomic “art” (Pet et al., 2018). Many of the doctors surveyed showed concern over how to address the findings from consumer genomic tests that told healthy people they were at risk for serious diseases. They worried that the results could lead to unnecessary medical treatment, increased costs, potential complications and problems with insurance. Overall, they wanted the patients to be provided with clearly communicated, actionable results: “What does it mean and what should I do?”

How actionable or reliable can genomic test results be? If we scan the brush-strokes of the whole genome, many of us (including medical people) would be left scratching our heads and saying, “What the…?” As we previously noted, genomes are noisy to the point of being called “space vomit.” Yet, there is still a belief that genetic testing will be pivotal to healthcare in the future (Pet et al., 2018).

We are taking steps in the right direction for learning how to improve the interpretation of genomic art in a meaningful way, but do we even need to? Novartis scientists recently conducted a study to link genetic variants (think “mutations”) to disease by using proteomics (Emilsson et al., 2018). In their analysis of nearly 5,000 proteins found in serum, they learned that many proteins clustered into groups that were co-regulated. The study also revealed that the protein groupings could be used to glean health insights.

What happened when genomic data were layered onto the picture? Well, the study reiterated the difficulty in linking a genetic variant to a single protein, which highlights the complexity of our bodies. However, the study suggested that some variants could be connected to the identified protein groups.

Maybe just focusing on and maximizing the more immediate potential of proteins to reveal the inner working of our bodies is the most direct and interpretable route for realizing the artistic promise of precision medicine.

 

References

Johnson, C. (2018, May 7). Medicine’s Wild West: 10 new genetic tests enter the market each day. The Washington Post. Retrieved on August 1, 2018 from https://www.washingtonpost.com/news/wonk/wp/2018/05/07/medicines-wild-west-10-new-genetic-tests-enter-the-market-each-day/?utm_term=.6cf729004c2e.

Emilsson, V., Ilkov, M., Lamb, J. R., Finkel, N., Gudmundsson, E. F., Pitts, R., . . . Gudnason, V. (2018). Co-regulatory networks of human serum proteins link genetics to disease. Science. doi:10.1126/science.aaq1327

Pet, D. B., Holm, I. A., Williams, J. L., Myers, M. F., Novak, L. L., Brothers, K. B., . . . Clayton, E. W. (2018). Physicians’ perspectives on receiving unsolicited genomic results. Genet Med. doi:10.1038/s41436-018-0047-z

 

Our Proteomic Fingerprint

Our Proteomic Fingerprint

Your genome has been hacked! …literally. Recently, a direct-to-consumer genetic test company was hacked and  information about millions of people was compromised (Thielking, 2018). In this day and age, it is not surprising when a company gets hacked. What is surprising is one of the ways a genetics company can be hacked. The sneaky route involves the malware being programmed into a DNA sample. When sequenced, the malware grants the hackers access. Sounds like a Hollywood movie plot, but it happened (Thielking, 2018).

Just how private is our biological information, and are people concerned? In a recent survey, about 47% people had concerns about the confidentiality of their genetic information derived from genealogical genetic testing (Hensley, 2018). Sadly, genetic information is difficult to keep secret. Aside from companies or other holders of genetic information being hacked, you can also lose your genetic privacy if your blood relatives choose to give away their genetic info like beads at Mardi Gras. This should not be a surprise, given the growing forensic use of relatives’ DNA to crack cold cases.

Is this only a genes problem, or are an individual’s proteins also a privacy risk? Although proteins can change rapidly in response to many different environmental changes, we can be named by our proteins. Not too long ago, a group of researchers used a mass spectrometry approach to identify a person based on the proteins found in a hair sample (Parker et al., 2016). They demonstrated that if a person carries a genetic variation that alters the amino acid sequence of a protein, then that person could be identified based on the protein sequence in the sample. Although hair has a relatively small number of proteins, it is likely that more complex protein samples, such as blood, could also be used to single out a person.

Is there a way to make mass spectrometry less revealing? One proposal is to remove some of the potentially distinguishing  data, which would make it harder to link it back to a specific individual, but not impossible (Li, Bandeira, Wang, & Tang, 2016). Nevertheless, this is a step in the right direction, but it may already be too late.

A quick search on Google reveals numerous repositories where proteomic data have been shared. When it comes time to publish a scientific paper, some journals, such as PLOS journals), make sharing proteomic data mandatory (http://journals.plos.org/plosone/s/data-availability). It is worth noting that a few exceptions to making the data available do exist, but a chance of the data holder being hacked still remains. Although some journals merely recommend sharing the data (e.g., Cell https://www.cell.com/cell/authors), there are growing cries for more data transparency, including (Matheson, 2018). The government has made a similar proposition (Friedman, 2018). With such a demand for transparency and access to data, can we still hold onto our beloved privacy? And how does this affect people’s willingness to donate biological samples or partake in clinical studies? The only thing that is certain now is that once the data are out, there is no way to secure them again.

 

References

Curran, A. M., Fogarty Draper, C., Scott-Boyer, M. P., Valsesia, A., Roche, H. M., Ryan, M. F., . . . Kaput, J. (2017). Sexual Dimorphism, Age, and Fat Mass Are Key Phenotypic Drivers of Proteomic Signatures. J Proteome Res, 16(11), 4122-4133. doi:10.1021/acs.jproteome.7b00501

Friedman, L. (2018, March 26) The E.P.A. Says It Wants Research Transparency. Scientists See an Attack on Science. New York Times. (Retrieved on June 6, 2018 from https://www.nytimes.com/2018/03/26/climate/epa-scientific-transparency-honest-act.html).

Hensley, S. (2018, June 1) POLL: Genealogical Curiosity Is A Top Reason For DNA Tests; Privacy A Concern. NPR. (Retrieved on June 3, 2018 from https://www.npr.org/sections/health-shots/2018/06/01/616126056/poll-genealogical-curiosity-is-a-top-reason-for-dna-tests-privacy-a-concern).

Li, S., Bandeira, N., Wang, X., & Tang, H. (2016). On the privacy risks of sharing clinical proteomics data. AMIA Jt Summits Transl Sci Proc, 2016, 122-131.

Matheson S. (2018, May 30) Why you should deposit your raw data. Crosstalk [blog post]. Retrieved on June 6, 2018 from http://crosstalk.cell.com/blog/why-you-should-deposit-your-raw-data).

Parker, G. J., Leppert, T., Anex, D. S., Hilmer, J. K., Matsunami, N., Baird, L., . . . Leppert, M. (2016). Demonstration of Protein-Based Human Identification Using the Hair Shaft Proteome. PLoS One, 11(9), e0160653. doi:10.1371/journal.pone.0160653

Thielking, M. (2018, June 5) Genealogy site MyHeritage says 92 million user accountscompromised. STAT. (Retrieved on June 6, 2018 from https://www.statnews.com/2018/06/05/genealogy-site-myheritage-says-92-million-user-accounts-compromised/).

 

Proteomics and the Warrior King

Proteomics and the Warrior King

The warrior king of Sparta, King Leonidas, and his small legion found themselves arrayed against an army of hundreds of thousands. Though they fought bravely until the very end, the odds and ultimately the Fates did not favor the poor king and his legion in that gory battle of Thermopylae.

Such battles are waged not just historically, but also today, though in much tinier, subcellular arenas. Consider Duchenne muscular dystrophy (DMD), a disease that predominantly affects young men. The origins of the battle start at the genetic level (National Institutes of Health (NIH), 2018). Mutations in the DMD gene, which can happen sporadically or be inherited on the X chromosome, compromise the critical protein dystrophin. According to the NIH’s war reports, dystrophin plays a primary role in stabilizing and guarding muscle fibers. Losing it leads to a constant attack on muscle tone and integrity, and ultimately to muscle defeat.

Further reading of the NIH’s war reports supplies added details. Early on in the battle, the brave young warriors bear the full brunt of the attack. Although their muscles struggle mightily to maintain themselves, children with DMD experience mobility problems early. Their calf muscles become enlarged as fat cells and other tissue replace the muscle. Over time, this muscle destruction becomes widespread, leading to breathing difficulties, joint problems and enlarged hearts (cardiomyopathy). In addition to compromised bodies, the battle also wreaks havoc on their minds, burdening the warriors with cognitive impairment, communication problems and impaired social behavior. Like the valiant King Leonidas and his army, these modern warriors are doomed.

But maybe we can delay or even change the fate of current or future warriors. Maybe there are additional weapons we have not considered. For example, in a piece of intel we learn that dystrophin is important not only to muscle formation, but also to signaling (i.e., communications between cells), and disruption of communications could be contributing to the disease (Allen, Whitehead, & Froehner, 2016). The intel’s authors postulate that a deeper understanding about the disruption in communications could open the door for more therapeutics as well as improvements in diagnosis, checking disease progression and assessing the effectiveness of treatments (Allen et al., 2016).

The SOMAscan platform, a proteomic technology, has been enlisted in battle against the tragic disease. The technology revealed that as few as six proteins maybe needed to accurately diagnose the disease (Parolo et al., 2018). Proteomics also revealed that DMD altered the communications in the immune system, the neurotrophin signaling pathway, apoptosis (cell death) and additional effectors of other biological systems (Parolo et al., 2018).

Once diagnosed, it becomes imperative to continue surveillance of the disease’s progression. One way to monitor progression involves invasive muscle biopsies. Recently, researchers developed a new less invasive way to monitor the progression using the SOMAscan platform, and it only requires a blood sample (Spitali et al., 2018). Over time, the tests revealed that the levels of hundreds of proteins changed, which could be indicative of muscle deterioration, increase of fat cells and heart problems (Spitali et al., 2018). Another group used the SOMAscan platform to identify biomarkers for cardiomyopathy in DMD patients (Anderson et al., 2017).

The war of the ages continues, but it may not last forever. As mentioned, proteomics can reveal the weaknesses of the DMD enemy and the effects of therapeutic strategies at the molecular level (Hathout et al., 2016). Let us hope that with different tactics, the fates may start to favor the brave and valiant warriors.

 

References

Allen, D. G., Whitehead, N. P., & Froehner, S. C. (2016). Absence of Dystrophin Disrupts Skeletal Muscle Signaling: Roles of Ca2+, Reactive Oxygen Species, and Nitric Oxide in the Development of Muscular Dystrophy. Physiol Rev, 96(1), 253-305. doi:10.1152/physrev.00007.2015

Anderson, J., Seol, H., Gordish-Dressman, H., Hathout, Y., Spurney, C. F., & Investigators, C. (2017). Interleukin 1 Receptor-Like 1 Protein (ST2) is a Potential Biomarker for Cardiomyopathy in Duchenne Muscular Dystrophy. Pediatr Cardiol, 38(8), 1606-1612. doi:10.1007/s00246-017-1703-9

Hathout, Y., Conklin, L. S., Seol, H., Gordish-Dressman, H., Brown, K. J., Morgenroth, L. P., . . . Hoffman, E. P. (2016). Serum pharmacodynamic biomarkers for chronic corticosteroid treatment of children. Sci Rep, 6, 31727. doi:10.1038/srep31727

National Institutes of Health. Duchenne muscular dystrophy. Retrieved on June 27, 2018 from https://rarediseases.info.nih.gov/diseases/6291/duchenne-muscular-dystrophy.

Parolo, S., Marchetti, L., Lauria, M., Misselbeck, K., Scott-Boyer, M. P., Caberlotto, L., & Priami, C. (2018). Combined use of protein biomarkers and network analysis unveils deregulated regulatory circuits in Duchenne muscular dystrophy. PLoS One, 13(3), e0194225. doi:10.1371/journal.pone.0194225

Spitali, P., Hettne, K., Tsonaka, R., Charrout, M., van den Bergen, J., Koeks, Z., . . . Aartsma-Rus, A. (2018). Tracking disease progression non-invasively in Duchenne and Becker muscular dystrophies. J Cachexia Sarcopenia Muscle. doi:10.1002/jcsm.12304

 

We Are All Mutants Here

We Are All Mutants Here

As most of you know, we are all mutants. Each of us carries variations in our gene sequence that, collectively, mark us uniquely. What does it mean to have a mutation? The simple answer is that you have a genetic sequence that is different than the decided upon consensus or “wild-type” sequence.

Well, what decides the “wild-type” sequence? To put it simply, it is the sequence most typically seen to date. This is where we enter a paradox. Pending the source of the genetic information used to decide the “wild-type” sequence, we could potentially be using information that is relevant for one demographic, but not for another.

The realization of this paradox is not a new phenomenon. For instance, Maynard Olson concluded that a single wild-type genetic sequence is a mere illusion and a wild-type human simply does not exist (Olson, 2011). He also said, “…genetics is unlikely to revolutionize medicine until we develop a better understanding of normal phenotypic variation (Olson, 2011).”

If we look at the literature, it seems that these words fell onto deaf ears or were just placed on a side burner. Since 2005, the number of studies involving genome-wide association studies (GWAS) to look for genetic mutations (i.e., variations) that can indicate a person’s risk of disease has exploded (Gallagher & Chen-Plotkin, 2018). Many associations have been made about genetic changes and a variety of diseases, however, they are only correlative in most cases. Compared to the deluge of GWAS studies, little has been done to determine that the associations found are indeed causing the disease (Gallagher & Chen-Plotkin, 2018).

Even if a direct link has been found, it does not completely explain why people who harbor genetic mutations known to cause detrimental disease appear perfectly healthy. In looking at a population of ~500,000 individuals, a recent genomic analysis revealed that 13 individuals harbored mutations that normally give rise to severe Mendellian childhood diseases but these people show no physical manifestations of the diseases (Chen et al., 2016). While this might seem a rare event, it really is not. From analyzing the genomes of 1000 “healthy” people, one group estimated that an average individual may actually be harboring >400 genetic changes that can damage the person’s biological equilibrium and >2 known disease causing mutations (Xue et al., 2012).

With everyone potentially harboring so many genetic changes that could have profound consequences, how is it that the clear majority of the people on this planet are functioning, and even functioning well? As James Evan from the University of North Carolina said in an NPR article, “The good news is that most of those mutations do not overtly cause disease, and we appear to have all kinds of redundancy and backup mechanisms to take care of that (Stein, 2012).”

What should we take away from all of this with regards to our individual health? Well, this is another reminder that our genes really only convey a risk and not an imminent fate. This is particularly true when the link between a genetic change and physical outcome is only correlative and not yet directly linked. Until researchers sift through all the associations found through GWAS to identify the ones that actually cause problems (this might not even be possible for the vast majority of associations found), we need to focus on the phenotypes. This includes – perhaps primarily – the proteins encoded by the genes and how they respond to our environment. Focusing on how proteins respond to environmental cues may begin to reveal the buffering systems and lead us to a path of health enlightenment.

 

References

Chen, R., Shi, L., Hakenberg, J., Naughton, B., Sklar, P., Zhang, J., . . . Friend, S. H. (2016). Analysis of 589,306 genomes identifies individuals resilient to severe Mendelian childhood diseases. Nat Biotechnol, 34(5), 531-538. doi:10.1038/nbt.3514

Gallagher, M. D., & Chen-Plotkin, A. S. (2018). The Post-GWAS Era: From Association to Function. Am J Hum Genet, 102(5), 717-730. doi:10.1016/j.ajhg.2018.04.002

Olson, M. V. (2011). Genome-sequencing anniversary. What does a “normal” human genome look like? Science, 331(6019), 872. doi:10.1126/science.1203236

Stein, R. (2012, December 6) Perfection Is Skin Deep: Everyone Has Flawed Genes. NPR. (Retrieved on May 16, 2018 from https://www.npr.org/sections/health-shots/2012/12/06/166648187/perfection-is-skin-deep-everyone-has-flawed-genes).

Xue, Y., Chen, Y., Ayub, Q., Huang, N., Ball, E. V., Mort, M., . . . Genomes Project, C. (2012). Deleterious- and disease-allele prevalence in healthy individuals: insights from current predictions, mutation databases, and population-scale resequencing. Am J Hum Genet, 91(6), 1022-1032. doi:10.1016/j.ajhg.2012.10.015

 

Why Dread the Very Thing That Once Brought Us so Much Joy?

“Happy birthday to you…” Do these words instill incapacitating fear or unbridled joy? Children welcome them with wide open arms and with great jubilation. As people age, the enthusiasm wanes to the point of dread. Why?

One possibility has to do with the aging process. Children look forward to getting older because with it come new found freedoms or rites of passage. At some point, we realize that aging is not as cool as we once thought. As we age, we fear the grim possibility of losing the very freedoms that we coveted in our youth, such as independence. But what if this did not have to be the case?

A key facilitator to our independence is our degree of health, which also decides how well we age. Those who are considered to be aging well typically look younger and are more active than one would expect for their chronological age. What if we have a laxer definition? What if the defining aspect was having a body that is biologically younger than what is stated on some document? For instance, a man may be 83 on paper but could really be 10 years younger based on how well his body works at the molecular level. It seems like science fiction, but it is already here in some ways. For example, a patient might be told by her doctor that her heart is functioning incredibly well for her age.

What we do not yet have is a clinical test that measures thousands of protein levels from many biological systems and breaks the news to us if they – and we – are aging well or not. But we are getting closer. In a first-of-its-kind study, researchers utilized the SOMAscan Platform to chart how our proteomes change with age and found proteins that tracked well with biological aging, which could lead to a better understanding of the molecular underpinnings of the process (Menni et al., 2015). Many of these proteins have been linked previously to aging, but it is unclear how others are contributing to the aging process. Although the results are incredibly promising, more and larger studies are needed to verify and expand on them.

Nevertheless, how exciting will it be to have a test that reveals our biological age? We could be significantly younger than what we are told by some calendar! It would certainly take the sting out of the next time someone wishes us a happy birthday. Who knows, birthdays might once again be a source of joy instead of dread.

 

References

Menni, C., Kiddle, S. J., Mangino, M., Vinuela, A., Psatha, M., Steves, C., . . . Valdes, A. M. (2015). Circulating Proteomic Signatures of Chronological Age. J Gerontol A Biol Sci Med Sci, 70(7), 809-816. doi:10.1093/gerona/glu121

 

Donuts Are Not the Only Things With Holes in Them

My love-hate nemesis is back. One encounter and our paths take a long time to diverge again. Despite all my efforts, my avoidance tactics have again failed and our paths have again converged. I speak, of course, of a delicious donut and its long-term clinging to my waist line.

As we age, our metabolism slows. The battle to shed pounds and support the svelte forms of our youth only becomes more difficult. But, numerous companies have sprung up to “help” us with our Herculean efforts to behave. Do any of them offer a valid method to help me sever ties permanently with the seductive donut?

Perhaps. Some genetic tests are now commercially available that promise to tell you the ideal diet based on your genes. Perusing some of the gene-recommended diets, it is clear they have captured what has always been promoted as a healthy diet (Robbins, 2016; Miller, 2018). Alas, I am looking for genetic permission to follow a diet rich in donuts and other sugary bakery goods. Nevertheless, people on the old-fashioned diet do spout off that they feel great and are losing weight (Miller, 2018). What gives? Could it truly be that genetics gave an insight or was it just following a healthy diet?

The successes (and not all people see success!) are not likely the result of properly matching diets to genotype. A recent and thorough study out of Stanford University, published in the Journal of the American Medical Association (JAMA), revealed that matching diets to genotypes does not give a person a huge advantage in weight loss (Gardner et al., 2018). Regardless of a match or not, participants lost weight. Eating healthy diets drove the weight loss. What this study does show is that we still have a lot to learn about the molecular underpinnings of weight loss and cannot rely on recommendations based solely on genetics.

If we look deeper into the biology of weight gain/loss, it is messier than my face after an encounter with delectable lemon curd filled donut. There are many reasons for the messiness; here are just a few:

1. People can lose weight based on the power of suggestion.

Yes, the power of suggestion is truly remarkable. In a study, two groups of women in the same profession were either told that their current level of activity met the surgeon general’s definition of an active lifestyle or were not given that info at all (Spiegel, 2008). After one month, the group informed about the surgeon general’s definition lost weight but had not increased their activity levels. The authors attributed the weight loss to a change in mindset. Quick! I need a doc or dietician to tell me that donuts can induce weight loss!

2. Genes can be turned on or off.

Research has shown that healthy lifestyle habits continually beat the genetic code predictions (Wang et al., 2018). It may sound unbelievable, but genes can be turned off or on before we are even born or in response to current environment (Youngson & Morris, 2013). Simple genetic tests may not be discerning which of these “make me fat now” genes are on or off.

3. The bacteria in our gut could be playing a role in weight loss.

In case you were not aware, our intestines are full of bacteria. In mice, researchers found that the guts of obese individuals were populated with certain types of bacteria (Ley et al., 2005). Soon after, it was learned that skinny mice could become obese if given the gut bacteria from fat mice (Turnbaugh et al., 2006). These early findings opened the flood gates to research looking at the contributions of gut bacteria in weight loss for humans. Guess what? Research shows that if our gut is populated by certain types of bacteria, it may give us an edge on losing weight (Hjorth et al., 2017; Youngson & Morris, 2013). There is still much to be learned, though, and some members of the research community are even calling these hopeful early findings into question (Begley, 2016).

It turns out that there is much to consider (and much unknown still) when it comes to factors that could affect our girth. Ugh! There must be a simpler way to help us better gauge our biological response to diets or even forecast if a diet will work.

Good news everybody! Proteomics (i.e., looking at the proteins in our body) may be the answer. Proteins (not genes) carry out the bulk of the metabolism work that happens at the molecular level. Already, technology has allowed us to look at the biological changes of least 5,000 proteins simultaneously. Initial work has shown promise in showing us how our bodies respond to diets (Oller Moreno et al., 2018) and indicated when a diet will be a huge failure (Thrush et al., 2017).

I may not be able to continue my love affair with the seductive donut as I get older without paying the consequences of increased mass. However, it is good to know that we may one day have the health insights necessary to better predict a diet where I might be able to have the occasional delectable gooey donut without a longer-term commitment to it.

 

References

Begley, S. (2016, September 22). Is the gut microbiome an important cause of obesity? STAT. Retrieved from https://www.statnews.com/2016/09/22/gut-microbiome-obesity.

Gardner, C. D., Trepanowski, J. F., Del Gobbo, L. C., Hauser, M. E., Rigdon, J., Ioannidis, J. P. A., . . . King, A. C. (2018). Effect of Low-Fat vs Low-Carbohydrate Diet on 12-Month Weight Loss in Overweight Adults and the Association With Genotype Pattern or Insulin Secretion: The DIETFITS Randomized Clinical Trial. JAMA, 319(7), 667-679. doi:10.1001/jama.2018.0245

Hjorth, M. F., Roager, H. M., Larsen, T. M., Poulsen, S. K., Licht, T. R., Bahl, M. I., . . . Astrup, A. (2017). Pre-treatment microbial Prevotella-to-Bacteroides ratio, determines body fat loss success during a 6-month randomized controlled diet intervention. Int J Obes (Lond). doi:10.1038/ijo.2017.220

Ley, R. E., Backhed, F., Turnbaugh, P., Lozupone, C. A., Knight, R. D., & Gordon, J. I. (2005). Obesity alters gut microbial ecology. Proc Natl Acad Sci U S A, 102(31), 11070-11075. doi:10.1073/pnas.0504978102

Miller A. (2018, January 16). Should You Take a Genetic Test to Find the Best Diet for You? U.S.News. Retrieved from https://health.usnews.com/wellness/articles/2018-01-16/should-you-take-a-genetic-test-to-find-the-best-diet-for-you.

Oller Moreno, S., Cominetti, O., Nunez Galindo, A., Irincheeva, I., Corthesy, J., Astrup, A., . . . Dayon, L. (2018). The differential plasma proteome of obese and overweight individuals undergoing a nutritional weight loss and maintenance intervention. Proteomics Clin Appl, 12(1). doi:10.1002/prca.201600150

Robbins, R. (2016, November 3). Genetic tests promised to help me achieve peak fitness. What I got was a fiasco. STAT. Retrieved from https://www.statnews.com/2016/11/03/genetic-testing-fitness-nutrition/.

Spiegel, A. (2008, January3). Hotel Maids Challenge the Placebo Effect. NPR. Retrieved from https://www.npr.org/templates/story/story.php?storyId=17792517.

Thrush, A. B., Antoun, G., Nikpay, M., Patten, D. A., DeVlugt, C., Mauger, J. F., . . . Harper, M. E. (2017). Diet-resistant obesity is characterized by a distinct plasma proteomic signature and impaired muscle fiber metabolism. Int J Obes (Lond). doi:10.1038/ijo.2017.286

Turnbaugh, P. J., Ley, R. E., Mahowald, M. A., Magrini, V., Mardis, E. R., & Gordon, J. I. (2006). An obesity-associated gut microbiome with increased capacity for energy harvest. Nature, 444(7122), 1027-1031. doi:10.1038/nature05414

Wang, T., Heianza, Y., Sun, D., Huang, T., Ma, W., Rimm, E. B., . . . Qi, L. (2018). Improving adherence to healthy dietary patterns, genetic risk, and long term weight gain: gene-diet interaction analysis in two prospective cohort studies. BMJ, 360, j5644. doi:10.1136/bmj.j5644

Youngson, N. A., & Morris, M. J. (2013). What obesity research tells us about epigenetic mechanisms. Philos Trans R Soc Lond B Biol Sci, 368(1609), 20110337. doi:10.1098/rstb.2011.0337