Blood molecular activity may provide insights into a person’s fitness levels and the biological processes that enhance physical performance. A collaboration between MIT, GE HealthCare, and the U.S. Military Academy at West Point has led to a computational model connecting numerous molecular signals with fitness indicators. This model could guide future research to enhance fitness training and expedite recovery from injuries or illnesses.<\/p>\n
The team studied over 50,000 biomarkers in 86 cadets at West Point preparing for a military contest. Their analysis identified molecular pathways that contribute to higher fitness levels. “We wanted to reduce 50,000 measurements to around 100 that are likely linked to fitness, going beyond mere statistical correlations,” explained Ernest Fraenkel, a professor at MIT.<\/p>\n
These biomarkers, detectable through blood tests, could provide athletes or patients with chronic conditions insights into areas to target for minimizing injury risk, speeding recovery, or enhancing performance. The study’s lead author is Azar Alizadeh from GE HealthCare, with Fraenkel and Luca Marinelli as senior authors, published in Communications Biology.<\/p>\n
To explore the genetic foundation of traits like height, scientists use genome-wide association studies (GWAS). However, physical fitness involves complex genetic, physiological, and environmental factors. The researchers aimed to identify some of these factors through a study with 86 West Point volunteers training for the Sandhurst Military Skills Competition.<\/p>\n
Throughout the three-month study, participants attended up to five sessions, providing blood samples before and after intense workouts. Other measurements included lean muscle mass and VO2 max. The researchers examined over 50,000 biomarkers from these samples, focusing on those predictive of performance in the Army Combat Fitness Test (ACFT).<\/p>\n
Instead of correlating fitness and biomarkers using a computational model, the small sample size would likely yield random correlations. Therefore, the researchers built a network model of marker interactions, identifying connections like signaling pathways or transcription factors activating gene sets.<\/p>\n
“We created a network resembling a city map, finding ‘neighborhoods’ that light up together,” noted Fraenkel. Marinelli added, “We leveraged network bioinformatics to develop a predictive model identifying biological circuits driving physical traits predictive of ACFT scores.”<\/p>\n
Using the PhenoMol model, the team identified over 100 biomarkers linked to ACFT performance. These predictions were more precise than models ignoring network connections and matched those based on VO2 and muscle mass measurements.<\/p>\n
The biomarkers identified by PhenoMol grouped into cellular pathways like blood coagulation and the complement cascade, aiding recovery from tissue injury and stress response. Another cluster involved the urea cycle, responsible for ammonia elimination from protein breakdown. Markers linked to mitochondrial function were also noted.<\/p>\n
Fraenkel aims to determine which markers reflect current fitness and which predict potential fitness levels, revealing strengths not captured by standard tests. This could benefit athletic training and recovery from injury or disease, or aging-related capacity loss.<\/p>\n
These molecular markers could also aid clinical trials assessing the efficacy of food supplements and fitness programs. The researchers aim to streamline testing by narrowing biomarkers to a few that can be measured using a single blood test method.<\/p>\n
This research was supported by DARPA, with the authors’ views not necessarily representing those of the agency.<\/p>\n