Pervasive effects of Wolbachia on host body mass and metabolic rate

Abstract

Metabolic rate is a widely variable trait that reflects organismal energy expenditure. At least half of all arthropod species carry maternally transmitted Wolbachia bacteria, but it is poorly understood how these endosymbionts affect host physiology and fitness. We tested how two central host traits, body mass and standard metabolic rate (SMR), are altered by Wolbachia for nine diverse Drosophila fly species and 13 co-occurring Wolbachia variants diverged roughly 150 million years. Interspecific trait variation was structured according to the host phylogeny, but within each host system, Wolbachia almost always altered body mass and/or SMR. Wolbachia often caused increases or decreases in body mass that, in some cases, also led to mass-dependent effects on whole-organism SMR. Other Wolbachia altered mass-adjusted SMR. These variable effects are plausibly explained by complex interactions among Wolbachia and host genomes, as documented for other traits (e.g., Wolbachia-induced cytoplasmic incompatibility). Endosymbiont effects on body mass and metabolic rate likely have cascading consequences for host biology and evolution, because these central traits are correlated with many aspects of host physiology, behavior, and life history. This also applies to hosts transinfected with Wolbachia from Drosophila to reduce the impacts of pathogens and pests on multiple continents.

INTRODUCTION

Metabolic rate is a measure of the energy that organisms expend to sustain life, integrating many aspects of organismal physiology, behavior, and life history. Metabolic rates vary widely among individuals and species (Burton et al. 2011; Konarzewski and Książek 2013; White and Kearney 2013; Auer et al. 2017), with implications for components of fitness, including growth, survival, behavior, and reproduction (Glazier 2005; Biro and Stamps 2010; Burton et al. 2011; Pettersen et al. 2016; Pettersen et al. 2018; Glazier 2022; White and Marshall 2023). Resting metabolic rate and other baseline measures of metabolism characterize the minimal energy expenditure required for self-maintenance (Stearns 1998; Burton et al. 2011; Pontzer and McGrosky 2022; White et al. 2022). In arthropods and other ectotherms, standard metabolic rate (SMR) represents the lowest rate of metabolism in an inactive, post-absorptive, nonreproductive adult, measured at a particular temperature. SMR is highly variable in insects and other arthropods (e.g., Addo-Bediako et al. 2002^; Nespolo et al. 2003^; Messamah et al. 2017^; Careau et al. 2019^; Somjee et al. 2021^; Leahy et al. 2025) and identifying the underlying sources of this variation is central to understanding broadscale patterns of energy expenditure.

A key source of variation in metabolic rate is body size, which generally scales hypometrically with metabolic rate, both within and among species (Addo-Bediako et al. 2002; Glazier 2005; Niven and Scharlemann 2005; Chown et al. 2007; White et al. 2007; Kolokotrones et al. 2010; White and Kearney 2011; White and Kearney 2013; White et al. 2019; Glazier 2022; Harrison et al. 2022); although, the relationship between body mass and metabolic rate is generally less strong for ectotherms (White et al. 2007). After accounting for body mass, both intrinsic factors (e.g., sex: Tomlinson and Phillips 2015^; Arnqvist et al. 2017^; Videlier et al. 2019; mitochondrial function: Speakman et al. 2004^; Tieleman et al. 2009^; Hoekstra et al. 2013^; Hoekstra et al. 2018^; Matoo et al. 2019) and extrinsic factors (e.g., temperature: Gillooly et al. 2001^; Addo-Bediako et al. 2002^; Nespolo et al. 2003^; Lachenicht et al. 2010^; Alton et al. 2017) contribute to mass-adjusted metabolic rates. Even after accounting for these factors and controlling for phylogeny (Addo-Bediako et al. 2002; Chown et al. 2007; Leahy et al. 2025), there remains much unexplained variation in metabolic rates (Burton et al. 2011; Videlier et al. 2019; Glazier 2022; Harrison et al. 2022; Leahy et al. 2025).

For insects and other arthropods, interactions with heritable microbes may contribute to metabolic-rate variation. Insects host a diversity of endosymbionts that alter central aspects of host biology and fitness (Moran et al. 2008; Werren et al. 2008; Feldhaar 2011; McFall-Ngai et al. 2013; McCutcheon et al. 2019; Hoffmann and Cooper 2024). Endosymbionts may require additional energy expenditures by hosts (Evans et al. 2009; Clavé et al. 2022), considering they interact with aspects of host nutrition (Koga et al. 2003; Brownlie et al. 2009; Kwak et al. 2025; Lindsey et al. 2025; Preuss et al. 2026), immune function (Oliver et al. 2003; Hedges et al. 2008; Teixeira et al. 2008; Jaenike et al. 2010; Nichols et al. 2021), and thermoregulation (Russell and Moran 2005; Brumin et al. 2011; Hague, Caldwell, et al. 2020). Endosymbionts are also maternally co-inherited with mitochondria in the cytoplasm (Ballad et al. 1996; Bright and Bulgheresi 2010; Richardson et al. 2012; Russell et al. 2019; Camus et al. 2022). Mitochondrial function and cytonuclear interactions with the nuclear genetic background impact SMR and other metabolic traits (e.g., Arnqvist et al. 2010^; Montooth et al. 2010^; Mossman et al. 2016); however, it is generally unknown how metabolic rate is affected when endosymbionts are present in the cytoplasm (but see Evans et al. 2009).

Here, we test the hypothesis that endosymbionts alter host metabolic rate. We examined variation in body mass and metabolic rates (SMR) in diverse Drosophila fly species and then tested how Wolbachia endosymbionts affect these central organismal traits. Maternally transmitted Wolbachia infect at least half of arthropod species (Weinert et al. 2015) and many insects (Werren et al. 2008; Zug and Hammerstein 2012). Wolbachia are usually facultative and not required for host survival (Hoffmann and Cooper 2024; Hoffmann and Cooper 2025), with population frequencies varying over space and time (Hoffmann et al. 1998; Kriesner et al. 2013; Hamm et al. 2014; Cooper et al. 2017; Hague, Mavengere, et al. 2020; Ravikanthachari et al. 2026). Body mass and metabolic rates are widely studied in arthropods, including in Drosophila (e.g., Addo-Bediako et al. 2002^; Chown et al. 2007^; Jumbo-Lucioni et al. 2010^; Messamah et al. 2017^; Videlier et al. 2019^; Videlier et al. 2021), but the effects of Wolbachia are generally not considered.

Wolbachia cells are found throughout host tissues, including gonads, fat body, gut, salivary glands, hemocytes, and Malpighian tubules (reviewed in Pietri et al., 2016). Wolbachia localization to reproductive tissues mediates maternal transmission (Russell et al. 2019; Porter and Sullivan 2023; Radousky et al. 2023; Hague et al. 2024) and effects on host reproduction (cytoplasmic incompatibility; Shropshire et al. 2020^; Shropshire et al. 2022); however, the physiological consequences of localization in somatic tissues are largely unknown (Hosokawa et al. 2010; Moriyama et al. 2015; Hague, Caldwell, et al. 2020; Porter and Sullivan 2023). Understanding how Wolbachia impact central organismal traits like metabolic rate is particularly important, because little is known about the fitness effects that underlie Wolbachia spread in nature (Hoffmann et al. 1990; Kriesner and Hoffmann 2018; Ross et al. 2019; Hoffmann and Cooper 2024; Graham et al. 2025); although effects on fecundity, virus blocking, and other traits remain candidates (Weeks et al. 2007; Brownlie et al. 2009; Cogni et al. 2021). It is also critical for improving Wolbachia-based biocontrol programs to prevent human diseases (Hoffmann et al. 2011; Walker et al. 2011; Ross et al. 2019; Utarini et al. 2021; Lenharo 2023; Velez et al. 2023) and agricultural pests (Nikolouli et al. 2018; Gong et al. 2020), where transinfected hosts experience fitness costs (reviewed in Ross et al. 2019).

We quantified body mass and SMR for 3,102 adult flies from nine different Drosophila species, with and without their naturally occurring Wolbachia strains, to test whether endosymbionts alter host metabolic rates. We examined the effects of 13 divergent Wolbachia variants (Figure 1), many of which have recently switched hosts via horizontal and introgressive transfer (Turelli et al. 2018; Cooper et al. 2019; Shropshire et al. 2026). We documented pervasive—but variable—effects of Wolbachia on body mass and metabolic rate in almost all Wolbachia-host systems assayed. In some cases, Wolbachia effects on body mass altered whole-organism SMR due to the scaling relationship between body mass and metabolic rate. Other Wolbachia variants altered mass-adjusted SMR after controlling for body mass. The widespread, variable effects we observed may be explained by novel cytonuclear interactions that have resulted from rapid Wolbachia host switching. Our findings imply that facultative endosymbionts are an important, unexplored source of variation in metabolic rates and related organismal traits.

Figure 1. Divergent Wolbachia strains found in Drosphila host species.

Estimated Bayesian phylogram of A- and B-group Wolbachia strains using 211 single copy genes of identical length spanning a total of 178,569 bp. The estimated divergence date with confidence intervals for A- and B-groups is reproduced from Shropshire et al. (2026). Naturally occurring Drosophila host species are listed in parentheses.

MATERIALS AND METHODS

Flies

We quantified body mass and SMR for Wolbachia-negative and -positive flies from nine different Drosophila species spanning roughly 22 million years of host evolution (Figure 2, Supplemental Table S1) (Suvorov et al. 2022; Shropshire et al. 2026). For two of the fly species, D. melanogaster and D. simulans, we tested multiple genotypes carrying different Wolbachia variants, which included wMel-D. melanogaster genotypes from northern and southern latitudes (Maine, USA and Panama City, Panama, respectively) and D. simulans lines carrying wRi, wHa, and a co-infection of wHa and the B-group wNo strain (Figure 1) (O’Neill and Karr 1990; Mercot et al. 1995; Rousset and Solignac 1995; James et al. 2002). In total, the dataset included 13 different Wolbachia-Drosophila host systems. First, we analyzed Wolbachia-negative flies from each genotype (N = 1,544) to quantify interspecific patterns of variation in body mass and SMR in a comparative framework, absent of any putative effects from Wolbachia. Next, we incorporated Wolbachia-positive individuals (N = 1,558) into focused analyses of each genotype to evaluate how Wolbachia alter body mass and SMR within each Wolbachia-host system (Supplemental Figures S1, S2).

Figure 2. Body mass and SMR estimates from Wolbachia-negative flies show strong evidence of phylogenetic signal.

(A) Estimated Bayesian phylogram of Drosophila host species using 20 conserved single-copy genes. (B) Mean body mass (± 95% CIs) of Wolbachia-negative flies for each sex of each host genotype. (C) Least square (LS) mean mass-unadjusted SMR (± 95% CIs) of Wolbachia-negative flies for each sex of each host genotype. LS mean values are extracted from genotype-specific LMMs that account for fly activity, experiment start time, flow rate in the chamber, water vapor, temperature, and light intensity. (D) LS mean mass-adjusted SMR (± 95% CIs) of Wolbachia-negative flies for each sex of each host genotype. LS mean values are extracted from genotype-specific LMMs that account for body mass, in addition to the nuisance variables listed in panel C. Wolbachia-negative data from the wRi-D. simulans genotype are shown for the D. simulans tip on the phylogeny.

In an earlier study, we analyzed effects of Wolbachia on locomotor activity of the flies in this study (Hague et al. 2021). As described in Hague et al. (2021), Wolbachia-positive flies were treated with tetracycline to generate paired Wolbachia-negative versions of each host genotype (Cooper and Shropshire 2024). The absence of Wolbachia was confirmed with qPCR. The gut microbiome of the tetracycline-cleared flies was then reconstituted by rearing them on food in which Wolbachia-positive males of the same genotype had fed and defecated for the prior 48 hours. Extra care was taken to avoid any potential detrimental effects of the antibiotic treatment by giving the flies a minimum of three generations after the tetracycline treatment before experiments (Ballard and Melvin 2007). Experiments took place many generations after the tetracycline treatment.

All flies were reared at 25°C under a 12L:12D light cycle (Percival Model I-36LL) on a standard food diet (Hague, Caldwell, et al. 2020; Hartman et al. 2024; Wheeler et al. 2024). Each day, we collected a batch of female and male virgins from the paired Wolbachia-negative and - positive lines of a single genotype. The four treatment groups (Wolbachia-negative females, - negative males, -positive females, -positive males) were maintained separately in isolation at 25°C for 3 to 5 days prior to flowthrough respirometry experiments.

Flowthrough respirometry

We measured metabolic rates of individual flies using a differential CO₂ analyzer (LI7000; Li-Cor Biosciences) and 16-chamber flowthrough respirometry and data acquisition system (MAVEn, Sable Systems International). The MAVEn incorporates a flow-distribution manifold, a main board (flow measurement, regulation, and control plus data acquisition and signal processing), and an activity board (sensors for activity, ambient temperature, humidity, and light intensity). The experiments were conducted at room temperature, and the readings from the activity board were used to account for any minor fluctuations in temperature, humidity, etc. in downstream analyses.

The CO₂ analyzer was calibrated weekly with CO₂-free air and 20 ppm CO₂ in N₂ (Norco). The baseline flow rate of gas was set to 35 ml min⁻¹. Approximately equivalent flow rates in the experimental channels (chambers 1–16) are maintained by means of matched flow resistances based on micro-orifice flow restrictors. During normal operation, a constant stream of dry CO₂-free is directed into the reference cell (A) of the CO₂ analyzer and then the airstream is humidified by directing it through Nafion tubing (Chemours) submerged in distilled water. The re-humidified airstream is then fed into the flow-distribution manifold, where it is split into 17 streams (one for each of the 16 animal chambers and one for the baseline). Each of the 16 animal chambers is a 2.4 ml volume tube. A mass flow meter on the MAVEn’s main board measures the actual flow rate of each airstream. Finally, the airflow is directed into the measurement cell (B) of the CO₂ analyzer, which records the fractional CO₂ concentration in excurrent air ($F_{E_{{CO}_2}}$).

Each day, we selected a batch of flies from a single genotype and randomly assigned the individual 3- to 5-day-old flies from each treatment group to one of the 16 animal chambers. The flies were gently aspirated into each chamber without anesthesia. Chambers were placed on the MAVEn activity board that uses infrared light (invisible to flies) to monitor animal activity in each chamber, sampled at 1 Hz. Metabolic measurements were performed between the hours of 09:00 and 16:00 during the day. We set the MAVEn to measure each chamber for 5 min (dwell time) with baseline measurements (5 min) taken every 2 chambers (i.e., interleave ratio = 2), resulting in a 120-min cycle to measure all 16 chambers. In each run, we allowed the system to complete two complete cycles, so that metabolic data for each fly was recorded twice (i.e., a minimum of 240 min). We monitored real-time flow rates to check for leaks, and chambers with erratic or unexpectedly low flows were removed prior to analysis. Immediately following each experiment, flies were anesthetized with CO₂ and their (wet) weights were measured using an analytical balance (Mettler Toledo). Chambers were washed with water and detergent after each run.

Data processing

Data were extracted and analyzed using R (R Core Team 2023). The CO₂ trace was corrected for drift using the interleaving baseline means. We modeled lines between the means of each interleaving baseline measurement and then, for each sampling timepoint, subtracted the baseline estimate from the actual measurement. We also corrected for the 3-sec lag between the activity channels and the CO₂ trace. The raw activity channels (one per chamber) were transformed into absolute difference sums (ADS). ADS was calculated by first calculating the cumulative sum of the absolute difference between consecutive activity readings and then calculating the slope of cumulative activity versus time (Videlier et al. 2019; Hague et al. 2021; Videlier et al. 2021). We discarded the initial 30 min of data from each run to avoid artifacts from the initial period during which flies adjust to the chambers.

Raw measures of CO₂ (ppm) were converted to molar rates of CO₂ production (${\dot{M}}_{{CO}_2}$) using the known flow rate (l min⁻¹) and the Ideal Gas Law:

$${\dot{M}}_{{CO}_2}=\frac{PFR{F}_{E_{{CO}_2}}}{RT}$$

where ${\dot{M}}_{{CO}_2}$ is the rate of CO₂ production of individual flies (mol min⁻¹), P is the pressure (1 atm), FR is the flow rate (l min⁻¹), $F_{E_{{CO}_2}}$ is the fractional CO₂ concentration in the excurrent air, R is the gas constant (0.08206 l atm K⁻¹ mol⁻¹), and T is temperature (K). For ease of viewing, ${\dot{M}}_{{CO}_2}$ values were converted to nmol min⁻¹ by multiplying by 10⁹.

The first 2 min of each metabolic measurement was discarded to allow the system to fully equilibrate after switching to a new chamber. From the remaining 3 min of each measurement period, we extracted the lowest 20-sec continuous bout of ${\dot{M}}_{{CO}_2}$ and calculated the average to generate an estimate of SMR. In addition to the average of the lowest 20-sec, we also extracted the start time of the SMR recording, the average flow rate (ml min⁻¹), water vapor pressure (ppt), temperature (°C), light intensity (lux), and activity (ADS) during the recording. While we previously found that Wolbachia can impact average activity levels of flies over a 3 hour period (Hague et al. 2021), our measurements of activity here over short 20-sec bouts were generally very close to zero for the majority of flies assayed (Supplemental Figure S3).

Phylogenetic analysis

We generated Wolbachia and Drosophila phylogenies to use in the phylogenetic comparative analyses described below. The Wolbachia phylogeny was generated in our previous analysis of host activity (Hague et al. 2021). The Bayesian phylogram of the A- and B-group Wolbachia was generated using 211 single-copy genes of identical length in all Wolbachia genomes, spanning a total of 178,569 bp (Figure 1). To generate a Drosophila host phylogeny (Figure 2), publicly available Drosophila sequences were obtained for the three wMel-D. melanogaster genotypes used in our experiments (Hague, Caldwell, et al. 2020; Hague et al. 2021), D. simulans (Hu et al., 2013), D. mauritiana (Meany et al. 2019), D. sechellia (Schrider et al. 2018), D. yakuba, D. santomea, D. teissieri (Cooper et al. 2019), D. suzukii (Turelli et al. 2018), and D. auraria (Conner et al. 2021). The host phylogeny was generated using the same 20 nuclear genes implemented in Turelli et al. (2018): aconitase, aldolase, bicoid, ebony, enolase, esc, g6pdh, glyp, glys, ninaE, pepck, pgi, pgm, pic, ptc, tpi, transaldolase, white, wingless, and yellow. We used BLAST with the D. melanogaster coding sequences to extract orthologs from the genomes of each host species. Sequences were then aligned with MAFFT 7. Finally, we used RevBayes and the GTR + Γ + I model partitioned by codon position and gene to accommodate potential variation in the substitution process among genes, as described in Turelli et al. (2018).

Analysis of body mass

We tested first for interspecific variation in body mass in a comparative framework, absent potential effects from Wolbachia. Body masses (mg) from Wolbachia-negative females and males of all genotypes were used in a linear model to test for effects of host genotype and sex on body mass with the “lm” function in R (R Core Team 2023). The significance of each effect was evaluated using F tests and type III sum of squares using the “Anova” function in the car R package (Fox and Weisberg 2019). We also used a phylogenetically aware model and the Drosophila phylogeny to test for genotype- and sex-differences in body mass while accounting for phylogenetic relatedness. Here, an extended phylogenetic generalized least squares (E-PGLS) regression was used with the “extended.pgls” function in the geomorph package (Collyer and Adams 2018; Baken et al. 2021; Collyer and Adams 2024; Adams et al. 2025). E-PGLS can accommodate more than one individual per species in a PGLS analysis (as opposed to using species means), which facilitates comparisons of within-species variation in body mass (i.e., females vs. males) while accounting for phylogenetic non-independence (Adams and Collyer 2024; Glynne and Adams 2026). Again, host genotype and sex were included as explanatory variables and the “anova” function was used to test for significance.

The body masses from Wolbachia-negative flies were also used to test for phylogenetic signal, the degree to which body mass variation is predicted by phylogenetic relationships. Mean body mass values for females and males of each genotype were used to test for phylogenetic signal on the Drosophila phylogeny using Pagel’s lambda (λ). A Pagel’s λ value of 0 coincides with character evolution that is independent of phylogenetic relationships, whereas λ = 1 is consistent with a Brownian motion model of trait evolution on the phylogeny. The “phylosig” function in the phytools package (Revell 2024) was used to compare fitted λ values to a model assuming no phylogenetic signal (λ = 0) using a likelihood ratio test. In this analysis, we used mean body mass estimates from Wolbachia-negative flies from the wRi-D. simulans genotype for the D. simulans tip on the Drosophila phylogeny; however, the means from the wHa or wHa-wNo D. simulans genotypes did not meaningfully alter the results.

To evaluate uncertainty surrounding estimates of λ, 95% confidence intervals were generated using 1,000 bootstrap replicates in the pmc package (Boettiger et al., 2012). When applicable, we conducted additional analyses to evaluate whether the number of taxa in the Drosophila phylogeny (N = 11 taxa) limited our ability to detect significant departures from λ = 0 (Hague, Caldwell, et al. 2020; Hague et al. 2021; Radousky et al. 2023). Small phylogenies are likely to generate near-zero values simply by chance, not necessarily because the phylogeny is unimportant for trait evolution (Boettiger et al. 2012). To evaluate whether larger phylogenies increased the accuracy of estimation, trees were simulated using an increasing number of taxa (N = 25, 50) and empirical λ estimates with the “sim.bdtree” and “sim.char” functions in the geiger package (Harmon et al. 2008). Confidence intervals were then re-estimated using the larger simulated trees.

Next, we incorporated Wolbachia-positive flies into focused analyses of each genotype to evaluate whether Wolbachia significantly alter female and/or male body mass within a given Wolbachia-host system. Here, the “lm” function was used to fit separate genotype-specific linear models for each Wolbachia-host system with Wolbachia infection status, sex, and a Wolbachia-by-sex interaction effect as explanatory variables. The significance of each effect was evaluated using F tests and type III sum of squares using the “Anova” function in the car R package (Fox and Weisberg 2019).

Finally, we used the Wolbachia phylogeny (Figure 1) to test if Wolbachia effects on host body mass show phylogenetic signal, addressing whether closely related Wolbachia strains tend to have similar effects on host body mass. Here, Wolbachia effects on host body mass were treated as a binary variable: each Wolbachia strain was scored based on whether it significantly affected host body mass in the genotype-specific linear models described above. We then tested for phylogenetic signal of the binary trait using the D statistic (Fritz and Purvis 2010), implemented in the caper package (Orme et al. 2018). The D statistic tests for significant departures from a Brownian expectation of phylogenetic signal (D = 0) and a model of phylogenetic randomness (D = 1).

Analysis of metabolic rate

We began the analysis of metabolic rate using the Wolbachia-negative flies to test whether mass-adjusted SMR varies interspecifically, absent potential effects from Wolbachia. To adjust for mass, body mass was included as an explanatory variable in the linear mixed model (LMM) analyses of SMR described below. Log-transformed SMR (nmol min⁻¹) values from the females and males of all genotypes were used in a LMM that included host genotype, sex, log-transformed body mass, and a sex-by-mass interaction effect as fixed effects, in addition to the following nuisance variables that could plausibly affect metabolic rate: average activity (ADS), start time of the SMR recording, average flow rate in the chamber (ml min⁻¹), average water vapor (ppt), average temperature (°C), and average light intensity (lux). The ID of each fly was included as a random effect to account for the fact that SMR of each fly was measured multiple times. The LMM was run using the “lmer” function in the lme4 package (Bates et al. 2015) and significance of the fixed effects was assessed using F tests and type III sum of squares with a Kenward-Roger approximation for degrees of freedom using the “Anova” function and the pbkrtest and car R packages (Halekoh and Højsgaard 2014; Fox and Weisberg 2019). We also used a phylogenetically aware model to test for genotype- and sex-differences in mass-adjusted metabolic rate while accounting for phylogenetic relatedness among Drosophila species using the “extended.pgls” function. For the E-PGLS, only the first SMR recording from each fly was included, because E-PGLS does not accommodate mixed models.

We also tested whether mass-adjusted SMR shows evidence of phylogenetic signal using the Drosophila phylogeny. Least-square (LS) mean mass-adjusted SMR values for Wolbachia-negative females and males were extracted from the genotype-specific LMMs described below (to account for the effects of body mass and nuisance variables on SMR) and used as mean trait values in tests for phylogenetic signal using Pagel’s λ. LS means were extracted using the “emmeans” function in the emmeans package (Lenth 2025). For the D. simulans tip on the Drosophila phylogeny, LS mean values from the wRi-D. simulans genotype were used; however, LS means from the wHa or wHa-wNo genotypes did not meaningfully alter the results.

Next, we incorporated Wolbachia-positive flies into focused analyses of each genotype to test whether Wolbachia significantly alter female and/or male metabolic rate within a given Wolbachia-host system. Here, LMMs were fit for each Wolbachia-host system that included the following fixed effects: Wolbachia infection status, sex, and a Wolbachia-by-sex interaction effect, in addition to the nuisance variables log-body mass, activity, start time, flow rate, water vapor, temperature, and light intensity. The ID of each fly was included as a random effect. F tests and type III sum of squares with a Kenward-Roger approximation for degrees of freedom were used to test the significant of fixed effects.

In the above-mentioned analysis of body mass, Wolbachia significantly altered body mass in the majority of Wolbachia-host systems (see Results, Figure 3). Metabolic rates usually scale with body mass, which we confirmed with interspecific and genotype-specific analyses of mass scaling relationships (see detailed analysis in Supplemental Materials). Thus, it is plausible that Wolbachia effects on body mass could indirectly alter SMR (e.g., Wolbachia increase host body mass, which increases whole-organism SMR as a consequence). We sought to distinguish between (1) mass-dependent effects of Wolbachia that alter whole-organism SMR and (2) mass-independent effects of Wolbachia that alter mass-adjusted SMR. Therefore, each of the genotype-specific LMMs was run excluding and then including body mass as an explanatory variable to test whether Wolbachia alter host metabolic rates before and after accounting for body mass.

Figure 3. Pervasive, variable effects of Wolbachia on host body mass and metabolic rate.

(A) Mean body mass (± 95% CIs) of Wolbachia-negative (W-) and -positive (W+) flies for each sex of each Wolbachia-Drosophila host system. (B) Mass-dependent effects of Wolbachia on SMR. LS mean mass-unadjusted SMR (± 95% CIs) of Wolbachia-negative and -positive flies for each sex of each genotype. (C) Mass-independent effects on SMR. LS mean mass-adjusted SMR (± 95% CIs) of Wolbachia-negative and -positive flies for each sex of each genotype. Below, the cladogram depicts evolutionary relationships among Wolbachia variants. Data for the co-infected wHa-wNo-D. simulans genotype are shown on the far right. Genotypes with a significant main effect of Wolbachia (*) or a Wolbachia-by-sex interaction (†) are marked to indicate significance (* = P < 0.05; ** = P < 0.01; *** = P < 0.001).

Finally, we used the Wolbachia phylogeny to test whether Wolbachia effects on metabolic rate show evidence of phylogenetic signal using the D statistic. As with body mass, Wolbachia effects on host SMR were treated as a binary variable: each Wolbachia strain was scored based on whether or not it had a significant effect on host SMR in the genotype-specific LMMs described above. We tested for evidence of phylogenetic signal for mass-dependent and - independent effects on SMR.

RESULTS

Body mass

The body masses of all Wolbachia-negative flies were first used to test whether Drosophila body mass varied interspecifically (Figure 2). In a linear model, Drosophila genotype (F_12,1530 = 146.41, P < 0.001) and sex (F_1,1530 = 2419, P < 0.001) significantly affected body mass. The sex effect reflected that females are generally larger than males (Figure 2). We also performed E-PGLS to test for variation in body mass while accounting for phylogenetic relationships among Drosophila species assuming a Brownian motion model of evolution. Drosophila genotype (F_12,1530 = 10.16, P = 0.001) and sex (F_1,1530 = 2683.12, P = 0.001) still had significant effects on body mass after accounting for phylogeny.

We also used mean body masses from Wolbachia-negative flies of each genotype to quantify the strength of phylogenetic signal on the Drosophila phylogeny using Pagel’s λ (Figure 2). We detected strong evidence of phylogenetic signal for both females (λ = 0.977, P = 0.005) and males (λ = 0.997, P = 0.004). This was true regardless of what genotype was included for the mean body mass estimates for the D. simulans tip on the tree (i.e., Wolbachia-negative D. simulans from the wRi, wHa, or wHa-wNo genotypes). Consistent with recent work, these results suggest that interspecific variation in Drosophila body mass for both sexes is strongly associated with phylogenetic relationships, but that significant variation still exists after accounting for phylogeny (Rader et al. 2026).

Next, Wolbachia-positive flies were incorporated into focused analyses of each Wolbachia-host system to evaluate how sex, Wolbachia, and sex-by-Wolbachia interactions impact host body mass using linear models (Figure 3, Supplemental Table S2). Sex had a large, significant effect on body mass in every model, again reflecting sexual size dimorphism. Wolbachia had a significant main effect on body mass for nine of the 13 Wolbachia-host systems. Wolbachia had a significant effect for the wMelCS-D. melanogaster genotype (F_1,456 = 22.944, P < 0.001), such that Wolbachia-positive flies had greater body mass than -negative flies (Figure 3). This was also true for the wSan-D. santomea (F_1,204 = 10.607, P = 0.001), wRi-D. simulans (F_1,125 = 46.428, P < 0.001), wSh-D. sechellia (F_1,206 = 4.222, P = 0.041), wMau-D. mauritiana (F_1,142 = 4.652, P = 0.033), and co-infected wHa-wNo-D. simulans genotypes (F_1,234 = 5.29, P = 0.022). We found the opposite pattern for three other genotypes, Panama wMel-D. melanogaster (F_1,157 = 4.546, P = 0.035), wYak-D. yakuba (F_1,160 = 4.518, P = 0.035), and wHa-D. simulans (F_1,350 = 5.026, P = 0.026), where Wolbachia-positive flies generally had a smaller body mass than -negative flies. There was no evidence for sex-by-Wolbachia interaction effects on body mass.

Finally, we used the Wolbachia phylogeny (Figure 1) to test for phylogenetic signal using the D statistic, assessing whether closely related Wolbachia tend to have similar effects on host body mass. The D value of 1.454 indicated overdisperson (D > 1) of effects on body mass across the tree, and a model of Brownian phylogenetic structure could be rejected (D = 0, P = 0.001). In contrast, a model of no phylogenetic structure could not be rejected (D = 1, P = 0.76), implying that Wolbachia effects on body mass do not show phylogenetic signal on the Wolbachia phylogeny.

Metabolic rate

The estimates of SMR from all Wolbachia-negative flies were first used to test whether Drosophila metabolic rates vary interspecifically (Figure 2). The fitted LMM showed a strong effect of Drosophila genotype (F_12,1577.2 = 18.515, P < 0.001), sex (F_1,1530 = 17.979, P < 0.001), body mass (F_1,1502.9 = 371.467, P < 0.001), and a sex-by-mass interaction effect (F_1,1534 = 371.467, P < 0.001). In addition, the following nuisance variables were also significant: temperature (F_1,2780 = 136.796, P < 0.001), start time (F_1,1597.1 = 1117.129, P < 0.001), flow rate (F_1,1603.1 = 110.169, P < 0.001), and activity (F_1,2231.1 = 418.828, P < 0.001). See Supplemental Table S3 for more information about nuisance variable effects. The E-PGLS incorporating Drosophila phylogenetic relationships indicated that Drosophila genotype no longer had a significant effect on SMR after accounting for phylogeny (F_12,1522 = 1.492, P = 0.105). In contrast, sex (F_1,1522 = 13.408, P = 0.001), body mass (F_1,1522 = 198.149, P = 0.001), and the sex-by-mass interaction effect (F_12,1522 = 4.87, P = 0.029) were still significant after accounting for phylogeny. Additionally, temperature (F_12,1522 = 58.564, P = 0.001), start time (F_12,1522 = 209.903, P = 0.001), flow rate (F_12,1522 = 400.586, P = 0.001), and activity (F_12,1522 = 138.4312, P = 0.001) were significant in the phylogenetically informed model.

We also used mean estimates of metabolic rate from Wolbachia-negative flies to characterize the strength of phylogenetic signal on the Drosophila phylogeny. Least-square (LS) mean mass-adjusted SMR values were extracted from the genotype-specific LMMs (described below) to test for phylogenetic signal using Pagel’s λ (Figure 2). LS mean mass-adjusted SMR values from females likely exhibit phylogenetic signal (λ = 0.926, P = 0.174), though the estimate was not significantly different from λ = 0. A power analysis indicated that the number of taxa in the Drosophila tree (N = 11) likely limited our ability to detect departures from λ = 0 (see Supplemental Figure S4). For males, there was strong, significant phylogenetic signal (λ = 0.947, P = 0.026). These results were generally similar regardless of the genotype we used for LS mean mass-adjusted SMR estimates for the D. simulans tip on the tree. Together, the comparative results from Wolbachia-negative flies indicate that much of the interspecific variation in mass-adjusted SMR can be explained by phylogenetic relationships, particularly for male flies.

Next, Wolbachia-positive flies were incorporated into focused analyses of each Wolbachia-host genotype to evaluate how sex, Wolbachia, and sex-by-Wolbachia interactions impact SMR using LMMs (Figure 3, Supplemental Tables S4, S5). Here, we compared LMM results before and after including body mass as an explanatory variable in order to distinguish Wolbachia effects on whole-organism SMR vs. mass-adjusted SMR. Sex had a significant main effect on SMR for all the genotypes in the mass-unadjusted models, reflecting that females generally have a higher whole-organism metabolic rate due to sexual size dimorphism. Mass-adjusted SMR varied by sex for only a few genotypes. Females had a greater mass-adjusted SMR than males for the Maine wMel-D. melanogaster (F _1,327.8 = 35.96, P < 0.001) and wHa-D. simulans (F _1,346.1 = 32.319, P < 0.001) genotypes, whereas females had a lower mass-adjusted SMR for the wYak-D. yakuba (F _1,163.3 = 5.229, P = 0.024) and wAur-D. auraria (F_1,177.8 = 10.986, P = 0.001) genotypes.

In a few Wolbachia-host systems, Wolbachia altered whole-organism SMR due to aforementioned Wolbachia effects on body mass and the mass scaling relationship with metabolic rate (see mass scaling analysis in Supplemental Materials). For example, Wolbachia increased SMR for the wMelCS-D. melanogaster genotype in the mass-unadjusted SMR model (F_1,452.4 = 11.352, P < 0.001); however, the effect disappeared in the mass-adjusted model (F_1,449.7 = 0.131, P = 0.718), implying that the increase in mass-unadjusted SMR is a result of wMelCS increases to host body mass (F_1,456 = 22.944, P < 0.001). After accounting for body mass, we no longer detected a Wolbachia effect on SMR. This was also true for the wRi-D. simulans genotype: Wolbachia increased SMR in the mass-unadjusted model (F_1,122.8 = 7.553, P = 0.007), but not in the mass-adjusted model (F_1,122.4 = 2.576, P = 0.111), implying that wRi increases in whole-organism SMR are due to increases in body mass (F_1,125 = 46.428, P < 0.001). For the Panama wMel-D. melanogaster genotype, Wolbachia decreased SMR in the mass-unadjusted model (F_1,153.9 = 6.101, P = 0.015), but not in the mass-adjusted model (F_1,152.7 = 3.037, P = 0.0834), indicating that the decrease in whole-organism SMR is due to the fact that Panama wMel decreases body mass (F_1,157 = 4.546, P = 0.035). For the wSuz-D. suzukii genotype, we also found evidence that Wolbachia decreases SMR in the mass-unadjusted model (F_1,154.9 = 5.78, P = 0.017); however, this effect disappeared in the mass-adjusted model (F_1,153.9 = 3.341, P = 0.07) despite the fact that wSuz does not seem to affect host body mass (F_1,158 = 1.243, P = 0.267).

For other systems, Wolbachia altered host metabolic rate in a mass-independent manner. For the Maine wMel-D. melanogaster genotype, Wolbachia increased SMR in both the mass-unadjusted (F_1,332.5 = 10.718, P = 0.001) and mass-adjusted models (F_1,329.6 = 7.277, P = 0.007). We found the opposite pattern for the wHa-D. simulans genotype, where Wolbachia decreased SMR in both the mass-unadjusted (F_1,345.9 = 9.406, P = 0.002) and mass-adjusted models (F_1,344 = 4.788, P = 0.029). For the wSan-D. santomea genotype, there was a significant sex-by-Wolbachia interaction effect for both the mass-unadjusted (F_1,196 = 7.566, P = 0.007) and mass-adjusted models (F_1,193.8 = 9.462, P = 0.002), wherein Wolbachia-positive females had a greater metabolic rate than -negative females, but males showed the opposite pattern. For the co-infected wHa-wNo-D. simulans genotype, we detected a negative effect of Wolbachia on SMR only after accounting for body mass in the mass-adjusted model (mass-unadjusted: F_1,230.8 = 1.305, P = 0.255; mass-adjusted: F_1,228.7 = 5.796, P = 0.017). These genotype-specific effects on SMR generally did not correlate with the Wolbachia effects on host locomotor activity (measured over a 3-hour period) that we documented for the flies in this study (Supplemental Figure S5; Hague et al. 2021).

Finally, we used the Wolbachia phylogeny (Figure 1) to test whether closely related Wolbachia strains tended to have similar effects on host metabolic rate. Using results from the mass-unadjusted LMMs, the estimate of D = 0.831 was high enough to reject a model of Brownian phylogenetic structure (D = 0, P = 0.024), but not a model of random phylogenetic structure (D = 1, P = 0.335). Similarly, the D = 1.422 value based on the mass-adjusted LMMs was high enough to reject a D = 0 model (P = 0.005), but not a D = 1 model (P = 0.773). Together, these results suggest that mass-dependent and -independent effects of Wolbachia on SMR do not show evidence of phylogenetic signal on the Wolbachia phylogeny.

DISCUSSION

Wolbachia had pervasive effects on the body mass and metabolic rate of Drosophila hosts. Interspecific variation was largely driven by host phylogenetic relationships, but within a given system, body mass and/or SMR were almost always impacted by Wolbachia. In some instances, Wolbachia altered whole-organism metabolic rates indirectly due to effects on body mass. In others, Wolbachia directly altered metabolic rates through mass-independent effects. This variation from one host system to the next is consistent with complex effects arising from interactions among Wolbachia and host genomes, as documented for other traits, including Wolbachia-induced cytoplasmic incompatibility. Below, we identify underlying mechanisms that may explain mass-dependent and -independent effects on SMR and then discuss the implications for endosymbiont-host relationships.

Mass-dependent effects

Wolbachia altered body mass for most hosts, including genotypes carrying A-group Wolbachia, the divergent B-group, and the wHa-wNo co-infection (Figure 3). In some instances, Wolbachia effects on body mass also altered whole-organism SMR. In most cases, we found that Wolbachia increased body mass, which is consistent with previous studies focused mostly on wMel in D. melanogaster. For example, wMel has been associated with increased female body mass (Kriesner et al. 2016) and, on nutrient-poor food, wMel increased pupae size and adult emergence rates (Lindsey et al. 2025). Outside of wMel-D. melanogaster, female D. innubila carrying wInn (another A-group strain) were slightly larger and showed higher fecundity when reared on nutrient-poor food (Unckless and Jaenike 2012; Hill et al. 2022).

Such findings have led to the hypothesis that facultative Wolbachia are involved in nutrient provisioning (Zug and Hammerstein 2015a; Newton and Rice 2020), a phenomenon observed in obligate Wolbachia relationships (Foster et al. 2005; Hosokawa et al. 2010; Nikoh et al. 2014). Positive effects on host growth could arise through several routes. For instance, genes associated with the heme biosynthetic pathway are generally found in Wolbachia genomes (Newton and Rice 2020), and wMel-positive D. melanogaster have increased fecundity when reared on iron-restricted or -overloaded diets (Brownlie et al. 2007; Brownlie et al. 2009). Wolbachia are also involved in synthesis of the B vitamins biotin and/or riboflavin in a number of host species (Hosokawa et al. 2010; Nikoh et al. 2014; Moriyama et al. 2015; Gerth and Bleidorn 2017; Preuss et al. 2026). Riboflavin is an essential cofactor for mitochondrial electron-transport-chain enzymes, and Wolbachia-synthesized riboflavin is linked to accelerated ovarian development and higher energy production in the rice planthopper Laodelphax striatellus carrying the wStri strain (group B) (Bing et al. 2020; Niu et al. 2025). Finally, removal of wMel from D. melanogaster exacerbated insulin/insulin growth factor-like signaling (IIS) mutant phenotypes, implying that wMel may increase IIS in Wolbachia-positive flies (Ikeya et al. 2009).

Certain Wolbachia-host systems (Panama wMel, wYak, and wHa) showed the opposite pattern, in which Wolbachia decreased body mass. These results highlight the complexity of our findings. Endosymbionts may increase growth under certain conditions, but they also likely consume host nutrients (Clavé et al. 2022). Wolbachia titer can change in response to host diet, implying that Wolbachia rely on nutrients from host cells (Ponton et al. 2015; Serbus et al. 2015). Increased food intake has been documented for wMel-positive males and, after 24 hours of fasting, males had lower body mass than did Wolbachia-negative flies, suggesting that wMel requires increased host energy expenditure (Zhang et al. 2021). Genomic and metabolomic analyses indicate that wMel interacts with a wide range of host metabolic pathways in D. melanogaster, including aspects of amino acid, carbohydrate, and lipid metabolism (Wu et al. 2004; Jiménez et al. 2019; Carneiro Dutra et al. 2020; Currin-Ross et al. 2021; Zhang et al. 2021; Lindsey et al. 2025). wMel may also depress IIS pathways (contrary to the results described above), which could help explain decreases in body mass (Haqshenas et al. 2019; Currin-Ross et al. 2021). It is also plausible that if Wolbachia impact the composition of host microbiota, this could have negative effects on body mass (Simhadri et al. 2017; Carneiro Dutra et al. 2020). Together, this work suggests that Wolbachia have widespread effects on body mass that likely depend on the context, including the physiological state of the host.

Mass-independent effects

We also detected mass-independent effects of Wolbachia on SMR. wMel increased mass-adjusted SMR for the Maine D. melanogaster genotype, whereas wHa and the wHa-wNo coinfection depressed it in D. simulans. For the wSan-D. santomea genotype, Wolbachia increased mass-adjusted SMR in females but decreased it in males. We hypothesize that these complicated mass-independent effects result from interactions between Wolbachia and mitochondria. Wolbachia are maternally co-transmitted with mitochondria (Henry and Newton 2018; Russell et al. 2019; Porter and Sullivan 2023; Radousky et al. 2023; Hague et al. 2024), and multiple lines of evidence imply that Wolbachia interact with mitochondria. For example, RNAi knockdown of multiple mitochondrial genes altered Wolbachia titer in Drosophila cell lines and developing oocytes (White et al. 2017), and wMel altered expression of host genes related to mitochondrial metabolism in D. melanogaster larvae (Lindsey et al. 2025). wMelCS also altered expression of glutamate dehydrogenase, a key enzyme in the TCA cycle, as well as ATP levels in D. melanogaster (Carneiro Dutra et al. 2020).

Wolbachia can also affect levels of reactive oxygen species (ROS) and the host oxidative environment, which may reflect Wolbachia interactions with the host immune system. These effects have been documented for wMelCS in D. melanogaster and wRi in D. simulans (Brennan et al. 2008; Brennan et al. 2012; Pan et al. 2012; Wong et al. 2015; Zug and Hammerstein 2015b; Monnin et al. 2016). Similarly, Wolbachia interactions with host iron metabolism could plausibly affect ROS levels and mitochondrial function (Kremer et al. 2009; Gill et al. 2014; Currin-Ross et al. 2021).

Complexity of host effects

We found that Wolbachia have widespread—but variable—effects on key physiological traits. Comparative studies of diverse Drosophila species have documented similar patterns of complex variation in the strength of Wolbachia-induced cytoplasmic incompatibility (Shropshire et al. 2022; Bagchi et al. 2026), maternal transmission fidelity (Radousky et al. 2023; Hague et al. 2024), and Wolbachia effects on host behavior (Hague, Caldwell, et al. 2020; Hague et al. 2021). We hypothesize that Wolbachia interactions with the host genomes underlie complex variation in host effects. For instance, the Maine and Panama wMel variants are very similar (zero third-position pairwise differences across 740 single copy genes, 757,926 total bp), yet differed in their effects on body mass and metabolic rate in their natural D. melanogaster backgrounds. Host backgrounds are known to modulate Wolbachia effects on host fecundity (Olsen et al. 2001), the strength of cytoplasmic incompatibility (Reynolds and Hoffmann 2002; Cooper et al. 2017), and maternal transmission rates (Hague et al. 2022). Wolbachia titers can also vary substantially across host backgrounds (Veneti et al. 2004; Mouton et al. 2007; White et al. 2017; Funkhouser-Jones et al. 2018; Grobler et al. 2018; Hague et al. 2022), and titer differences in somatic tissues from one host system to the next could have varying effects on host physiology. Extending our analyses to host genotypes with reciprocally introgressed nuclear and cytoplasmic genomes (e.g., Hague et al. 2020^; Hague et al. 2022) is required to distinguish nuclear and cytoplasmic contributions to Wolbachia effects on body mass and metabolic rate.

Recent host switching may also help explain the variable patterns we observed. Wolbachia, including variants in our study, are known to rapidly switch between divergent host species (O’Neill et al. 1992; Raychoudhury et al. 2009; Ahmed et al. 2015; Gerth and Bleidorn 2017; Turelli et al. 2018; Cooper et al. 2019; Shropshire et al. 2026). For instance, wMel and related “wMel-like” variants (Figure 1) diverged only 10⁵–10⁶ years ago, yet occur in hosts diverged 350 MYA spanning holometabolous insects (Shropshire et al. 2026). wRi and related variants—including wSuz and wAur—have spread even more rapidly across divergent Drosophila (Turelli et al. 2018), and the wHa, wSech, wNo, and wMau strains found within the D. simulans host clade (Figure 2) have histories inconsistent with cladogenic acquisition (Ballard 2000; Meany et al. 2019). Host switching generates novel genomic combinations that could plausibly affect mitochondrial function, energy production, and interactions with the nuclear genome. Cytonuclear interactions between mitochondria and the nuclear background are known to affect metabolism (Arnqvist et al. 2010; Montooth et al. 2010; Mossman et al. 2016), and future work should examinine how Wolbachia impact these interactions. Somatic Wolbachia injections can be used to decouple the association between Wolbachia and host mitochondria in the cytoplasm (Frydman et al. 2006; Frydman 2007), which will be necessary for dissecting effects on host metabolism. Tripartite interactions among Wolbachia, mitochondria, and the nuclear genome are also likely to influence the physiology of transinfected hosts. Understanding these effects has the potential to improve the application efficacy of variants like wMel that block dengue and other arboviruses in mosquitoes (Hoffmann et al. 2011; Walker et al. 2011; Ross et al. 2019).

Conclusions

Patterns of Wolbachia spread in natural host populations indicate that Wolbachia confer a fitness benefit to hosts, yet little is known about such benefits in nature (Weeks et al. 2007; Hamm et al. 2014; Cooper et al. 2017; Meany et al. 2019; Ross et al. 2019; Hague, Mavengere, et al. 2020; Cogni et al. 2021; Hague et al. 2022; Hoffmann and Cooper 2024; Graham et al. 2025). Wolbachia effects on body mass and metabolic rate, two central host traits, are likely to have cascading effects on aspects of host fitness, including growth and survival (Chown and Gaston 2010; Burton et al. 2011; White and Kearney 2013). Body mass effects in particular are likely to impact other aspects of host biology, because so many physiological and ecological traits are correlated with body size (Brown et al. 2004; Chown and Gaston 2010; White and Kearney 2011). Wolbachia effects on SMR—a measure of the cost of self-maintenance—are also likely to have wide-ranging implications for hosts. Ultimately, these metabolic effects contribute to the overall impact of Wolbachia on host fitness, which determines endosymbiont spread and persistence in host populations (Hoffmann et al. 1990; Kriesner et al. 2016; Hoffmann and Cooper 2024; Graham et al. 2025).

Our findings indicate that future studies involving arthropod body mass, metabolism, and related life-history traits should consider the presence of facultative endosymbionts. A growing body of work shows that facultative Wolbachia alter a range of important host traits with fitness implications, including fecundity (Hoffmann et al. 1990; Olsen et al. 2001; Fry et al. 2004; Weeks et al. 2007), nutrition (Brownlie et al. 2009; Hosokawa et al. 2010; Nikoh et al. 2014; Newton and Rice 2020; Lindsey et al. 2025), immune function (Hedges et al. 2008; Teixeira et al. 2008; Osborne et al. 2009; Martinez et al. 2014; Cogni et al. 2021; Bruner-Montero and Jiggins 2023; Porter and Sullivan 2023), temperature preference (Arnold et al. 2019; Truitt et al. 2019; Hague, Caldwell, et al. 2020), and activity levels and sleep (Vale and Jardine 2015; Bi et al. 2018; Morioka et al. 2018; Hague et al. 2021). Elucidating how Wolbachia affect host traits like body mass and metabolic rate is critical for understanding host-endosymbiont evolution in natural systems and for optimizing endosymbiont-based applications in transinfected vector and pest systems (Hoffmann and Cooper 2025).

Funding Statement

This work was supported by National Science Foundation (NSF) CAREER (2145195) and National Institutes of Health MIRA (R35GM124701) Awards to BSC. Additional support was provided by the University of Montana Genomics Core (UMGC) and Montana INBRE Data Science Core, which are funded by the National Institute of General Medical Sciences (P20GM103474), the Office of the Vice President for Research and Creative Scholarship at the University of Montana, and the M. J. Murdock Charitable Trust (202324717 to BSC).

DATA AVAILABILITY

All data and code are available at https://github.com/jmgraham30/Metabolism_final_R_scripts.