Assessing Genetic Risk for Physical Activity and Its Interaction with Diet in Predicting Activity Levels and Weight Loss in the iMPROVE Study.

The physical activity polygenic score PGS002254 showed a nominally significant interaction with diet group for weight loss post-intervention.

• Interaction coefficient B = 7.57, SE = 3.57 × 10^0, p = 0.04
• Model explained R² = 0.06 of variance in log-transformed weight loss from baseline to month 3
• Weight loss was operationalized as log-transformed weight change from baseline to month 3
• Analysis was conducted in 197 participants with available genetic data from the iMPROVE study

The sedentary behavior polygenic score PGS001923 showed a significant interaction with a 'High in unsaturated fats and fruit juice consumption' dietary pattern for baseline MET-mins/week.

• Interaction coefficient B = 1.51 × 10³, SE = 4.135 × 10², p = 0.001
• Model explained R² = 0.091 of variance in baseline MET-mins/week
• The dietary pattern was previously extracted from the iMPROVE cohort
• Physical activity was operationalized as metabolic equivalent of task minutes per week (MET-mins/week) derived from self-reported activity questionnaires

The iMPROVE study sample consisted of 202 participants with baseline phenotypic data, predominantly classified as having obesity.

• Sample included 59 men and 143 women aged 19–65 years
• 75 participants were classified as having overweight and 126 as having obesity based on baseline BMI
• 197 of the 202 participants had available genetic data for PRS calculation
• Correlation analyses and linear regression models were used to assess main effects of PRSs and dietary patterns as well as gene-diet interactions

Dietary patterns moderated the association between genetic predisposition and both physical activity levels and weight loss outcomes.

• Interactions were examined using both randomized intervention dietary groups and previously extracted dietary patterns from the iMPROVE cohort
• The randomized intervention dietary groups were used as a moderator for the PGS002254–weight loss interaction
• Previously extracted dietary patterns were used as moderators for the PGS001923–MET-mins/week interaction
• Variance explained by models was modest (R² = 0.06 and R² = 0.091), underscoring the multifactorial nature of lifestyle behaviors

Weight-related outcomes examined included log-transformed weight loss from baseline to month 3 and change in BMI post-intervention.

• The intervention period assessed for weight outcomes extended to month 3
• BMI change post-intervention was included as a secondary weight-related outcome
• Log transformation was applied to weight loss values prior to analysis
• Both main effects of PRSs and gene-diet interactions were evaluated using linear regression models