Despite a plethora of interventions, including diet, exercise, and medications, obesity continues to be a significant health problem. The fight against obesity currently involves substantial trial-and-error, out-of-pocket costs, and adverse effects. Obesity medications cause gastrointestinal side effects in nearly 73% of patients.

Molecular predictors of response to obesity treatment are key in combating the obesity epidemic. Toward this end, the National Institutes of Health (NIH) launched ADOPT (Accumulating Data to Optimally Predict Obesity Treatment Core Measures) in 2018 which collects behavioral, biological, environmental, and psychosocial data to identify predictors of response to obesity treatment.1

Variables for weight loss

A clinical trial2 conducted at the Mayo Clinic by a team led by Andres Acosta, MD, and Michael Camilleri, MD, stratified obesity into four distinct observable traits or phenotypic categories based on underlying pathophysiology, and prescribed anti-obesity medications—phentermine, phentermine/topiramate, bupropion/naltrexone, lorcaserin, and liraglutide—according to the patient’s phenotype. The study conducted on 450 participants measured body composition, resting energy expenditure, satiety, eating behavior, and physical activity, among other parameters, using standard assays and questionnaires. Based on these parameters, the investigators were able to assign one of the four phenotypes to 383 of the 450 participants.

The phenotype-guided approach was linked to a 1.75-fold greater weight loss after a year, and the fraction of patients who lost more than 10% body weight at the end of the year was 79% compared to only 34% among patients who were prescribed medications without phenotyping.

Acosta and Camilleri went on to found Phenomix Sciences, a precision obesity biotechnology startup that aims to predict obesity phenotypes to improve outcomes of weight loss treatments in patients. With financial backing from the American Medical Association’s innovation arm, Health2047, and United Healthcare, Phenomix is developing a first-of-its-kind therapy selection blood test (Phenomix Sciences Obesity Platform) that predicts a patient’s response to GLP-1 (Glucagon-like peptide 1) treatments, such as Wegovy (semaglutide). The test is based on a proprietary biobanking registry that includes data collected from over 1,000 patients treated for obesity at the Mayo Clinic.

Precise phenotyping will change the future of obesity treatment believes Mark Bagnall, CEO at Phenomix Sciences.

Mark Bagnall, CEO of Phenomix Sciences, said, “This is an exciting time in the evolution of obesity medicine. The justifiable excitement about new weight management products such as semaglutide and tirzepatide has highlighted the value of rigorous research in the field.”

Phenotyping finesse

The team led by Acosta and Camilleri measures how patients respond on treadmill tests, their body composition, and gut mobility, among other things. “You get fed a radio-labeled omelet and neuroimaging tells you how quickly the omelet passes through your GI tract,” said Bagnall.

Based on detailed patient data on genetics, metabolomics, hormones, and behavior, the therapy selection test phenotypes patients into four obesity categories. These include hungry brain (satiation), hungry gut (satiety), emotional hunger (reward), and slow burn (metabolism). The satiation phenotype is categorized by the excessive consumption of calories without feeling full, while the satiety phenotype is marked by feeling hungry shortly after eating. Eating in response to emotional triggers is the hallmark of the emotional hunger phenotype, and a slow metabolism where the body burns calories ineffectively, indicates the slow burn phenotype.

Whereas clinical trials can conduct comprehensive phenotyping studies on a significant sample, pre-pandemic data estimates 41.9% of adults in the United States are obese.3 The pandemic might have worsened this statistic. Such a large patient population creates obvious challenges in scaling the ability to predict treatment responses through individual studies.

“The problem is you can’t send 100 million people to the Mayo Clinic for eight hours of observational analysis,” said Bagnall. “Our job as a company is to translate those observations into a blood test. Our test analyzes about 2,000 SNPs based on which we determine the four primary obesity phenotypes.”

Refining predictions

Bagnall believes that the company’s partnership with the Mayo Clinic biobanking study will continue to generate more data that will help the company refine its algorithm and the predictive set of SNPs for a better understanding of obesity and a more accurate phenotyping of patients from diverse populations.

Data analysts know the predictive value of large datasets is only as good as the data fed into the database. Clinical data can be messy and certain data points may not be available for all patients, requiring computational corrections for missing data. Rigorous quality control measures are therefore key for a functional database that enables accurate predictions.

“The big picture is to work with individuals and institutions who know what they are doing, whom you can trust. Part of the reason we work with Mayo Clinic is their level of rigor when it comes to onboarding patients and documenting their background,” said Bagnall. “Then we have internal controls on our side to make sure that we can pick out anomalies in the data we are receiving. There are checklists at both ends. So, we are able to quickly ensure that data is properly dovetailing.”

Precision medicine approaches recognize that obesity may result from a range of pathophysiological mechanisms. Therefore, not all patients respond to the same treatment. By enabling access to clinical and molecular data linked to stages and types of obesity, the biobanking registry helps evaluate variability in obesity treatment responses using a data-driven approach that provides personalized treatment options for patients.

“Our biobanking agreement with Mayo Clinic is an important opportunity to make vast strides in how we understand the complexities of obesity treatment. We believe the biobanking registry investment will better support obesity centers by providing concrete evidence and insights into how DNA and other factors need to be considered in treatment,” said Bagnall.

Age, race, gender, education, socioeconomic status, and behavior compound genetic factors that underlie obesity. Accurate phenotyping can pinpoint a patient’s specific cause of weight gain and match the patient to the optimal treatment option for their phenotype.

“The upside is significant for patients and payers. Patients get the right treatment the first time and payers avoid paying for a costly trial-and-error approach,” said Bagnall. “There are still physicians who believe that obesity is more a moral failing than a true disease. That is something that we in the healthcare community need to fix.”

Obesity was designated a disease by the American Medical Association in 2013. This has begun to dissolve the misconceptions surrounding the chronic, multifactorial disease. In addition to establishing clear phenotypes, phenotype-genotype correlations, and molecular mechanisms, it is also important to recognize the role of genotypic and phenotypic diversity in the various manifestations of obesity.

Phenotype-driven precision medicine holds the potential to change the future of obesity treatment through better-informed clinical decisions, determining coverage, and effectively combining diagnostic testing with drug therapy.


  1. Rosenbaum M, Agurs-Collins T, Bray MS, et al. Accumulating Data to Optimally Predict obesity Treatment (ADOPT): recommendations from the biological domain. Obesity (Silver Spring) 2018;26(suppl 2):S25-S34.
  2. Acosta A, Camilleri M, Abu Dayyeh B, et al. Selection of Anti-obesity Medications Based on Phenotypes Enhances Weight Loss: A Pragmatic Trial in an Obesity Clinic. Obesity (Silver Spring). 2021;29(4):662-671
  3. Stierman, B, Afful, J, Carroll, MD et al. National Health and Nutrition Examination Survey 2017–March 2020 Prepandemic Data Files Development of Files and Prevalence Estimates for Selected Health Outcomes. National Health Statistics Reports. Series : NHSR No. 158.
Previous articleWhite House Seeks Input on the National Biotechnology and Biomanufacturing Initiative
Next articleBreast Cancer Treatment Response Prediction Improved by Cellular Method