Practical aspects of estimating energy components in rodents

Area of Science:

• Metabolic physiology
• Computational biology
• Statistical modeling

Background:

• Indirect calorimetry generates complex data on energy expenditure.
• Component analysis aims to dissect total energy expenditure into resting metabolic rate (RMR), activity, and thermic effects.
• Accurate component analysis requires robust statistical methods and careful experimental design.

Purpose of the Study:

• To review and compare computational methods for indirect calorimetry data analysis.
• To investigate the impact of data resolution and system parameters on component analysis accuracy.
• To identify the most robust methods for dissecting energy expenditure.

Main Methods:

• Review of Kalman filtering, linear regression, penalized spline (P-spline) regression, and minimal energy expenditure estimation.
• Analysis of high- and low-resolution datasets from rodent indirect calorimetry.
• Investigation of factors affecting estimation accuracy: sample time, activity sensor accuracy, and chamber washout time.

Main Results:

• Strong correlation between Kalman filtering and P-spline regression for high-resolution data, except for activity respiratory quotient (RQ).
• P-spline regression accurately estimated basal metabolic rate (BMR) and resting RQ from low-resolution data (9 min sample time).
• Thermic effect of food (TEF) and activity-related energy expenditure (AEE) were sensitive to reduced sample rates.

Conclusions:

• Both Kalman filtering and P-spline regression are suitable for component analysis of continuous, single-channel indirect calorimetry data.
• P-spline regression provides more robust results for low-resolution data from multi-channel systems.
• Method selection should consider data resolution and system configuration for accurate metabolic component analysis.