MIT unveils PhenoMol fitness predictor

Physical fitness is hard to reduce to one number because endurance, strength, recovery, and metabolism all shape performance. MIT and its partners say a new model called PhenoMol can read some of that complexity from blood. (news.mit.edu) (nature.com) The idea is to look past surface measures like pace or heart rate and ask what cells are doing underneath. Researchers tracked patterns in genes, proteins, and metabolites — small molecules involved in energy use and recovery — to connect biology with fitness outcomes. (nature.com) (news.mit.edu) MIT, GE HealthCare, and the U.S. Military Academy at West Point unveiled the work in an MIT News report published April 28, after the underlying paper appeared in Communications Biology on February 17. Lead author Azar Alizadeh is a principal scientist at GE HealthCare, and MIT professor Ernest Fraenkel and GE HealthCare scientist Luca Marinelli are the senior authors. (news.mit.edu) (nature.com) To build the model, the team studied 86 West Point cadets preparing for the Sandhurst Military Skills Competition. They measured more than 50,000 biomarkers from blood samples and used prior biological knowledge to compress that data into what the paper calls “expression circuits,” or groups of linked signals that may reflect real pathways rather than random correlations. (news.mit.edu) (nature.com) That matters because fitness studies often have far more biological measurements than participants, which can produce false patterns. The Communications Biology paper says PhenoMol outperformed regression models that did not use its network-based dimensionality reduction. (nature.com) The researchers say they were trying to cut tens of thousands of measurements down to about 100 markers with a plausible mechanistic link to fitness. In the MIT report, Fraenkel said the goal was to move beyond “just a statistical correlation” toward markers with “a likelihood that there is a causal relationship.” (news.mit.edu) The paper frames the system as a research tool, not a consumer fitness product. The authors say it could be used in future studies of wellness, performance, and disease, while the MIT report says blood-based markers might eventually help tailor training, reduce injury risk, or speed recovery from chronic illness or long-term injury. (nature.com) (news.mit.edu) The backdrop is a broader push to map how exercise changes the body at the molecular level. The National Institutes of Health-backed Molecular Transducers of Physical Activity Consortium, or MoTrPAC, is also building a molecular map of physical activity across human studies. (motrpac.org) (nature.com) For now, PhenoMol is a way of asking whether a blood sample can capture part of what coaches and clinicians now infer from tests, wearables, and observation. MIT’s version of the answer is cautious: the signals are promising, but the next step is validation in broader groups beyond 86 cadets. (news.mit.edu) (nature.com)