# Exploring different definitions of methane phenotypes in Dutch Holstein cattle

## TL;DR

This study compares different ways to measure methane emissions in dairy cows and finds that some methods are genetically linked, which could help breed cows that produce less methane.

## Contribution

The paper evaluates multiple methane phenotype definitions and their genetic correlations in dairy cattle to identify suitable traits for genomic selection.

## Key findings

- Methane concentration and production phenotypes show high genetic correlations except for IPCC Tier2.
- Methane intensity has negative genetic correlations with most methane phenotypes.
- Average methane concentration and eructation peaks are suitable proxies for methane emissions.

## Abstract

Summary: Several methane phenotypes have been proposed in recent years. When measuring methane using breath analyzers (sniffers), 2 commonly used phenotypes are methane concentration (CH4c; ppm) and methane production (CH4p; g/d). However, even within these phenotypes, different definitions exist. For example, for CH4c, the average per visit is commonly used, but other phenotypes involving the identification of “eructation” peaks have been proposed. For CH4p, there are several available formulas to calculate it, using methane concentration or ratio as input, together with other production traits. Additionally, methane intensity is another phenotype that has interest, as it represents the amount of methane per production unit (in dairy cattle, milk). Genetic correlations within each group of phenotypes (CH4c and CH4p) were highly positive, except for the formula of the Intergovernmental Panel on Climate Change (IPCC) to predict methane that was used as a benchmark (Tier2). Methane intensity had negative genetic correlations with the majority of the phenotypes.

Summary: Several methane phenotypes have been proposed in recent years. When measuring methane using breath analyzers (sniffers), 2 commonly used phenotypes are methane concentration (CH4c; ppm) and methane production (CH4p; g/d). However, even within these phenotypes, different definitions exist. For example, for CH4c, the average per visit is commonly used, but other phenotypes involving the identification of “eructation” peaks have been proposed. For CH4p, there are several available formulas to calculate it, using methane concentration or ratio as input, together with other production traits. Additionally, methane intensity is another phenotype that has interest, as it represents the amount of methane per production unit (in dairy cattle, milk). Genetic correlations within each group of phenotypes (CH4c and CH4p) were highly positive, except for the formula of the Intergovernmental Panel on Climate Change (IPCC) to predict methane that was used as a benchmark (Tier2). Methane intensity had negative genetic correlations with the majority of the phenotypes.

•Average CH4c is positively correlated with most of the phenotypes except for ratio (CH4c and CO2 concentration).•Methane production phenotypes derived by formulas are highly positively correlated among them.•Methane intensity is positively correlated with most methane phenotypes.

Average CH4c is positively correlated with most of the phenotypes except for ratio (CH4c and CO2 concentration).

Methane production phenotypes derived by formulas are highly positively correlated among them.

Methane intensity is positively correlated with most methane phenotypes.

One of the most promising strategies to permanently reduce methane emissions in dairy cattle is through genomic selection, where the primary goal is to identify and selectively breed low-emitting ruminants. An important step is to define which trait definition to use. Several methane phenotypes have been proposed in recent years. When measuring methane using breath analyzers (sniffers), 2 commonly used phenotypes are methane concentration (CH4c; ppm) and methane production (CH4p; g/d). However, different definitions exist for both phenotypes. For example, for CH4c, the average per visit is commonly used, but other phenotypes involving the identification of eructation peaks have been suggested. Several formulas are available to calculate CH4p, using CH4c or the ratio between CH4c and CO2 concentrations as input, together with other production traits. Additionally, methane intensity (MeI), another phenotype of interest, is the amount of methane produced per milk unit (kg). Therefore, the aims of this study were (1) to estimate genetic parameters for 11 distinct phenotypes, including 3 CH4c definitions, carbon dioxide concentration per visit (CO2c), the ratio between CH4c and CO2c, 5 phenotypes for CH4p based on different formulas (Madsen, Chagunda, 2 of Kjeldsen, and IPCC Tier2 as a benchmark), and MeI; and (2) to estimate genetic correlations between these methane phenotypes and milk yield (MY) and BW. A total of 149,726 sniffer (CH4c and CO2c) records were available from 7,600 Dutch Holstein cows measured between 2019 and 2024 on 68 farms. Data were analyzed with an animal repeatability model with fixed effects, including herd-year-season, week of lactation, and lactation number with age of cow at calving nested. Estimated heritability values ranged from 0.16 to 0.30 for CH4c phenotypes, with the number of eructation peaks having the highest heritability. For CH4p, heritabilities ranged from 0.03 to 0.27, with Tier2 having the highest value. Heritabilities for the CH4c/CO2c ratio and CO2c were 0.08 and 0.13, respectively. Genetic correlations between CH4c phenotypes were moderate to highly positive, ranging between 0.49 and 0.85. Likewise, highly positive genetic correlations (between 0.89 and 1) were estimated for CH4p phenotypes, except for Tier2, which presented correlations between 0.04 and 0.37 with the other CH4p phenotypes. Average CH4c (avg) per visit and sum of maximum peaks of CH4c (speaks) had moderate to high positive correlations (0.36–0.95) with the majority of the CH4p phenotypes (except Tier2). Methane intensity had moderate to high positive genetic correlations (0.38 to 0.80) with the majority of the phenotypes except for Kjeldsen2 (0.03) and Tier2 (−0.90). Milk yield had positive genetic correlations with all the methane phenotypes (0.04 to 0.94) except for MeI (−0.66) and Kjeldsen2 (−0.69). Body weight had close to zero genetic correlations with CH4c phenotypes (−0.09 to 0.07), and a moderate positive to moderate negative correlation (−0.72 to 0.57) for CH4p phenotypes. Given their strong correlations with the other methane phenotypes, close to zero correlation with body weight, and no induced dependencies with BW and MY, as seen with the CH4p phenotypes, avg and speaks appear to suitable proxies for methane emissions when using sniffers.