PMC:4264897 / 8561-19256 JSONTXT

{"project":"MyTest","denotations":[{"id":"25512844-16573844-26483235","span":{"begin":5293,"end":5297},"obj":"16573844"},{"id":"25512844-23637808-26483236","span":{"begin":7129,"end":7133},"obj":"23637808"},{"id":"25512844-23637808-26483237","span":{"begin":7281,"end":7285},"obj":"23637808"},{"id":"25512844-23637808-26483238","span":{"begin":7627,"end":7632},"obj":"23637808"},{"id":"25512844-23637808-26483239","span":{"begin":7663,"end":7667},"obj":"23637808"},{"id":"25512844-23637808-26483240","span":{"begin":7854,"end":7858},"obj":"23637808"},{"id":"25512844-23637808-26483241","span":{"begin":8152,"end":8156},"obj":"23637808"}],"namespaces":[{"prefix":"_base","uri":"https://www.uniprot.org/uniprot/testbase"},{"prefix":"UniProtKB","uri":"https://www.uniprot.org/uniprot/"},{"prefix":"uniprot","uri":"https://www.uniprot.org/uniprotkb/"}],"text":"Methods and Materials\nLarge populations were established in the laboratory for G. texensis and A. domesticus. These populations comprised animals that were raised from hatching under identical and optimal environmental conditions and thus gave true estimates of the mass–length relationship (hereafter referred to as reference populations; Peig and Green 2010). These reference populations were used to determine correlations among mass components (absolute and relative mass) and body size measures, determine mass–length scaling relationships (i.e., b SMA, see below), and to examine how mass components correlate with condition scores from residual analysis and the scaled mass index.\nThe reference data set for G. texensis (n  = 86 males; n  = 103 females) comprised lab-reared descendants of crickets collected in Austin, TX (U.S.A.). These crickets were raised communally for their first 3 weeks in large bins (64 L) and then housed individually in 250-mL containers (10 cm diameter × 4.5 cm depth) until eclosion to adulthood. The reference data set for A. domesticus (n  = 59 males; n  = 61 females) comprised the crickets used in Worthington et al. (2013). Juvenile crickets (4–5 weeks of age) were acquired from a commercial dealer (Fluker's Cricket Farms, Port Allen, LA), and the sexes were separated prior to their imaginal molt, with females housed in large communal bins (44 × 33 × 40 cm) and males housed individually in 250-mL containers (10 cm diameter × 4.5 cm depth). For both species, containers were cleaned weekly and all individuals were provided with cotton-plugged water vials and fed dry cat food (Special Kitty: 34% protein, 13% fat) ad libitum. Crickets were reared and maintained at 27 ± 1°C on a 12 h:12 h light:dark cycle and were checked daily for eclosion to adulthood.\nAdult crickets were euthanized by freezing at −20°C either at eclosion (G. texensis) or 12–14 days post-eclosion (A. domesticus). Body mass (g) and pronotum length (mm) were recorded immediately after death in both species with the lengths of the left and right tibia and femura also being recorded in G. texensis. Pronotum length was defined as the distance between the anterior and posterior edges of the pronotum. Pronota, tibia and femura were measured to the nearest 0.01 mm under a stereomicroscope using Leica LAS image analysis software (Leica Microsystems Inc., Buffalo Grove, IL).\nCrickets were dried at 60°C for 24 h and weighed to the nearest 0.01 mg using an electronic balance (Denver Instruments TP-64). Water mass was measured as the difference between fresh mass and dry mass. Body fat was then extracted using petroleum ether (Fisher Scientific, Hanover Park, Illinois, USA) reflux in a Soxhlet apparatus for 12 h. Individuals were again dried at 60°C for 24 h and then reweighed to obtain their lean dry mass. Body fat content (mg) was obtained by subtracting lean dry mass from dry mass.\nFor both species, we used Pearson product-moment correlation (r) to assess the strength of the relationship of absolute and relative mass components with body size measures for males and females separately as well as both sexes pooled. Relative (%) mass components were calculated as mass component divided by size measure multiplied by 100.\n\nBody condition validation\nBody condition at eclosion was calculated for each individual using Peig and Green's (2009) scaled mass index. This index standardizes body mass to a specific fixed value of a linear body measurement based on the scaling relationship between mass and length using the equation:\n(1)\nwhere M i and L i are the body mass and linear body measurement of individual i, respectively, b SMA is the scaling exponent estimated by the standardized major axis (SMA) regression of ln M on ln L; L 0 is an arbitrary value of L (e.g., the arithmetic mean value for the study population); and is the predicted body mass for individual i when the linear body measure is standardized to L 0.\nIn our G. texensis reference population, log body mass is positively correlated with log femur length (r  =   0.6467, P  \u003c   0.0001, n  = 189) and log tibia length (r  =   0.5635, P  \u003c   0.0001, n  = 189), but is most strongly correlated with log pronotum length (r  =   0.8326, P  \u003c   0.0001, n  = 189). In our A. domesticus reference population, log body mass is also strongly positively correlated with log pronotum length (r  =   0.7332, P  \u003c   0.0001, n  = 121). Therefore, pronotum length is an excellent linear indicator of body size in both cricket species and was used as L in our calculations of (G. texensis: L 0 = 3.349 mm; A. domesticus: L 0 = 2.908 mm).\nFor each of our reference populations, we first used model II regression to calculate the allometric slope (b SMA) of the best-fit line from a standardized major axis regression of fresh body mass on pronotum length (both variables log-transformed). The scaling mass index is superior to other methods of determining body condition from mass and length estimates because its use of model II linear regression (i.e., standardized major axis regression, henceforth SMA). SMA is superior to other regression techniques when, for example, both variables have some underlying error rate associated with their measurement and are measured on different scales (Warton et al. 2006; Peig and Green 2009). The model II slopes did not differ between the sexes in either species (see Results and Discussion), and so a common slope (G. texensis: b SMA = 2.642; A. domesticus: b SMA = 2.549) was calculated for each species. For each species, we calculated each individual's by substituting the appropriate slope and mean pronotum length (G. texensis: L 0 = 3.349 mm; A. domesticus: L 0 = 2.908 mm) into Eq. (1) along with each individual's fresh body mass (M i) and pronotum length (L i).\nWe used the same method to standardize the other body components (i.e., fat content, dry mass, lean dry mass, and water content) for a fixed size (M i in Eq. 1 was taken to be the mass of the component). Peig and Green (2009) recommended such standardization because body components (e.g., fat, protein, water, etc.) are generally correlated with body size. We note that the same L 0 value (i.e., pronotum length for G. texensis: L 0 = 3.349 mm; A. domesticus: L 0 = 2.908 mm) was used for both the scaled component mass and the scaled body mass index.\nResidual index (R i), for each cricket in the reference population, was calculated by entering log body mass as the dependent variable into an ordinary least squares (OLS) regression model with log pronotum length as the independent variable. The standardized residual was then extracted for each cricket. Prior to computing the common slope, we also tested whether the slopes differed between the sexes and diet treatments. Separate analyses of covariance were used to test whether the elevations (i.e., adjusted means) of the slopes differed between the sexes and diets.\nWe correlated scaled mass components with and R i for females, males, and both sexes pooled using Pearson product-moment correlation (r).\n\nComparing methods\nWe used data presented in Kelly and Tawes (2013) to compare the performance of the scaled mass index with that of two other commonly used approaches (residual index and ANCOVA). Kelly and Tawes (2013) examined the interaction between nutritional quality (poor vs. good diet) during development and sex on various fitness-related traits, including body condition, at adulthood. This data set comprised information on the body size (pronotum length, mm) and body mass (g) at eclosion for n  = 82 females and n  = 92 males (see Kelly and Tawes 2013 for details).\nKelly and Tawes (2013) calculated condition scores using , but the allometric scaling exponent they used was calculated from that data set (b SMA = 2.319). In this study, we calculated for Kelly and Tawes’ (2013) crickets using the b SMA from the G. texensis reference population (b SMA = 2.642; see above), while the mean pronotum length was the same in both cases (i.e., L 0   =   3.073 mm). As discussed by Peig and Green (2010), using the b SMA from the experimental population (e.g., Kelly and Tawes 2013) might not be ideal because the development of the test animals was manipulated via diet restriction and thus they might not exhibit the “true” scaling relationship. We substituted these values along with individual body mass (M i) and pronotum length (L i) into Eq. (1) to calculate for each cricket. Prior to calculating the common standard major axis regression slope (b SMA) for use in Eq. (1), however, we first tested the assumption that the slopes did not differ between the sexes and diet treatments by adding either sex or diet to the model and inspecting the interaction term (a significant interaction suggests that the slopes are heterogeneous). Similarly, we tested whether the elevations of standard major axis slopes differed within each factor by inspecting their 95% confidence intervals; the hypothesis of different elevations is rejected by overlapping confidence intervals. All variables were log-transformed prior to analysis. We tested the effect of sex and diet on by entering both of these fixed factors as independent variables into an ANOVA.\nWe calculated R i by entering the dependent variable log body mass into an OLS regression model with log pronotum length as the independent variable. The standardized residual was then extracted for each cricket. Prior to computing the common slope (b OLS) we tested whether the slopes differed between the sexes and diet treatments. Separate analyses of covariance were used to test whether the elevations (i.e., adjusted means) of the slopes differed between the sexes and diets. R i was then entered into an ANOVA as the dependent variable with sex and diet treatment entered as fixed-factor treatment variables.\nWe assessed the performance of ANCOVA by entering log-transformed body mass into a general linear model as the dependent variable with sex and diet treatment as fixed-factor independent variables and log pronotum length as a covariate. This procedure first required testing the homogeneity of slopes assumption; if the interactions between sex and log pronotum length, and diet and log pronotum length were statistically non-significant they were removed and the ANCOVA performed.\nFor our analyses of and R i, we tested for homoscedascity among the condition scores within each treatment using Levene's test. All statistical analyses were performed in R 3.0.2 (R Development Core Team 2014) using the packages lmodel2 (Legendre 2013), smatr (Warton et al. 2011), and ggplot2 (Wickham 2009). All statistical tests were conducted at the α  = 0.05 level."}

{"project":"2_test","denotations":[{"id":"25512844-16573844-26483235","span":{"begin":5293,"end":5297},"obj":"16573844"},{"id":"25512844-23637808-26483236","span":{"begin":7129,"end":7133},"obj":"23637808"},{"id":"25512844-23637808-26483237","span":{"begin":7281,"end":7285},"obj":"23637808"},{"id":"25512844-23637808-26483238","span":{"begin":7627,"end":7631},"obj":"23637808"},{"id":"25512844-23637808-26483239","span":{"begin":7663,"end":7667},"obj":"23637808"},{"id":"25512844-23637808-26483240","span":{"begin":7854,"end":7858},"obj":"23637808"},{"id":"25512844-23637808-26483241","span":{"begin":8152,"end":8156},"obj":"23637808"}],"text":"Methods and Materials\nLarge populations were established in the laboratory for G. texensis and A. domesticus. These populations comprised animals that were raised from hatching under identical and optimal environmental conditions and thus gave true estimates of the mass–length relationship (hereafter referred to as reference populations; Peig and Green 2010). These reference populations were used to determine correlations among mass components (absolute and relative mass) and body size measures, determine mass–length scaling relationships (i.e., b SMA, see below), and to examine how mass components correlate with condition scores from residual analysis and the scaled mass index.\nThe reference data set for G. texensis (n  = 86 males; n  = 103 females) comprised lab-reared descendants of crickets collected in Austin, TX (U.S.A.). These crickets were raised communally for their first 3 weeks in large bins (64 L) and then housed individually in 250-mL containers (10 cm diameter × 4.5 cm depth) until eclosion to adulthood. The reference data set for A. domesticus (n  = 59 males; n  = 61 females) comprised the crickets used in Worthington et al. (2013). Juvenile crickets (4–5 weeks of age) were acquired from a commercial dealer (Fluker's Cricket Farms, Port Allen, LA), and the sexes were separated prior to their imaginal molt, with females housed in large communal bins (44 × 33 × 40 cm) and males housed individually in 250-mL containers (10 cm diameter × 4.5 cm depth). For both species, containers were cleaned weekly and all individuals were provided with cotton-plugged water vials and fed dry cat food (Special Kitty: 34% protein, 13% fat) ad libitum. Crickets were reared and maintained at 27 ± 1°C on a 12 h:12 h light:dark cycle and were checked daily for eclosion to adulthood.\nAdult crickets were euthanized by freezing at −20°C either at eclosion (G. texensis) or 12–14 days post-eclosion (A. domesticus). Body mass (g) and pronotum length (mm) were recorded immediately after death in both species with the lengths of the left and right tibia and femura also being recorded in G. texensis. Pronotum length was defined as the distance between the anterior and posterior edges of the pronotum. Pronota, tibia and femura were measured to the nearest 0.01 mm under a stereomicroscope using Leica LAS image analysis software (Leica Microsystems Inc., Buffalo Grove, IL).\nCrickets were dried at 60°C for 24 h and weighed to the nearest 0.01 mg using an electronic balance (Denver Instruments TP-64). Water mass was measured as the difference between fresh mass and dry mass. Body fat was then extracted using petroleum ether (Fisher Scientific, Hanover Park, Illinois, USA) reflux in a Soxhlet apparatus for 12 h. Individuals were again dried at 60°C for 24 h and then reweighed to obtain their lean dry mass. Body fat content (mg) was obtained by subtracting lean dry mass from dry mass.\nFor both species, we used Pearson product-moment correlation (r) to assess the strength of the relationship of absolute and relative mass components with body size measures for males and females separately as well as both sexes pooled. Relative (%) mass components were calculated as mass component divided by size measure multiplied by 100.\n\nBody condition validation\nBody condition at eclosion was calculated for each individual using Peig and Green's (2009) scaled mass index. This index standardizes body mass to a specific fixed value of a linear body measurement based on the scaling relationship between mass and length using the equation:\n(1)\nwhere M i and L i are the body mass and linear body measurement of individual i, respectively, b SMA is the scaling exponent estimated by the standardized major axis (SMA) regression of ln M on ln L; L 0 is an arbitrary value of L (e.g., the arithmetic mean value for the study population); and is the predicted body mass for individual i when the linear body measure is standardized to L 0.\nIn our G. texensis reference population, log body mass is positively correlated with log femur length (r  =   0.6467, P  \u003c   0.0001, n  = 189) and log tibia length (r  =   0.5635, P  \u003c   0.0001, n  = 189), but is most strongly correlated with log pronotum length (r  =   0.8326, P  \u003c   0.0001, n  = 189). In our A. domesticus reference population, log body mass is also strongly positively correlated with log pronotum length (r  =   0.7332, P  \u003c   0.0001, n  = 121). Therefore, pronotum length is an excellent linear indicator of body size in both cricket species and was used as L in our calculations of (G. texensis: L 0 = 3.349 mm; A. domesticus: L 0 = 2.908 mm).\nFor each of our reference populations, we first used model II regression to calculate the allometric slope (b SMA) of the best-fit line from a standardized major axis regression of fresh body mass on pronotum length (both variables log-transformed). The scaling mass index is superior to other methods of determining body condition from mass and length estimates because its use of model II linear regression (i.e., standardized major axis regression, henceforth SMA). SMA is superior to other regression techniques when, for example, both variables have some underlying error rate associated with their measurement and are measured on different scales (Warton et al. 2006; Peig and Green 2009). The model II slopes did not differ between the sexes in either species (see Results and Discussion), and so a common slope (G. texensis: b SMA = 2.642; A. domesticus: b SMA = 2.549) was calculated for each species. For each species, we calculated each individual's by substituting the appropriate slope and mean pronotum length (G. texensis: L 0 = 3.349 mm; A. domesticus: L 0 = 2.908 mm) into Eq. (1) along with each individual's fresh body mass (M i) and pronotum length (L i).\nWe used the same method to standardize the other body components (i.e., fat content, dry mass, lean dry mass, and water content) for a fixed size (M i in Eq. 1 was taken to be the mass of the component). Peig and Green (2009) recommended such standardization because body components (e.g., fat, protein, water, etc.) are generally correlated with body size. We note that the same L 0 value (i.e., pronotum length for G. texensis: L 0 = 3.349 mm; A. domesticus: L 0 = 2.908 mm) was used for both the scaled component mass and the scaled body mass index.\nResidual index (R i), for each cricket in the reference population, was calculated by entering log body mass as the dependent variable into an ordinary least squares (OLS) regression model with log pronotum length as the independent variable. The standardized residual was then extracted for each cricket. Prior to computing the common slope, we also tested whether the slopes differed between the sexes and diet treatments. Separate analyses of covariance were used to test whether the elevations (i.e., adjusted means) of the slopes differed between the sexes and diets.\nWe correlated scaled mass components with and R i for females, males, and both sexes pooled using Pearson product-moment correlation (r).\n\nComparing methods\nWe used data presented in Kelly and Tawes (2013) to compare the performance of the scaled mass index with that of two other commonly used approaches (residual index and ANCOVA). Kelly and Tawes (2013) examined the interaction between nutritional quality (poor vs. good diet) during development and sex on various fitness-related traits, including body condition, at adulthood. This data set comprised information on the body size (pronotum length, mm) and body mass (g) at eclosion for n  = 82 females and n  = 92 males (see Kelly and Tawes 2013 for details).\nKelly and Tawes (2013) calculated condition scores using , but the allometric scaling exponent they used was calculated from that data set (b SMA = 2.319). In this study, we calculated for Kelly and Tawes’ (2013) crickets using the b SMA from the G. texensis reference population (b SMA = 2.642; see above), while the mean pronotum length was the same in both cases (i.e., L 0   =   3.073 mm). As discussed by Peig and Green (2010), using the b SMA from the experimental population (e.g., Kelly and Tawes 2013) might not be ideal because the development of the test animals was manipulated via diet restriction and thus they might not exhibit the “true” scaling relationship. We substituted these values along with individual body mass (M i) and pronotum length (L i) into Eq. (1) to calculate for each cricket. Prior to calculating the common standard major axis regression slope (b SMA) for use in Eq. (1), however, we first tested the assumption that the slopes did not differ between the sexes and diet treatments by adding either sex or diet to the model and inspecting the interaction term (a significant interaction suggests that the slopes are heterogeneous). Similarly, we tested whether the elevations of standard major axis slopes differed within each factor by inspecting their 95% confidence intervals; the hypothesis of different elevations is rejected by overlapping confidence intervals. All variables were log-transformed prior to analysis. We tested the effect of sex and diet on by entering both of these fixed factors as independent variables into an ANOVA.\nWe calculated R i by entering the dependent variable log body mass into an OLS regression model with log pronotum length as the independent variable. The standardized residual was then extracted for each cricket. Prior to computing the common slope (b OLS) we tested whether the slopes differed between the sexes and diet treatments. Separate analyses of covariance were used to test whether the elevations (i.e., adjusted means) of the slopes differed between the sexes and diets. R i was then entered into an ANOVA as the dependent variable with sex and diet treatment entered as fixed-factor treatment variables.\nWe assessed the performance of ANCOVA by entering log-transformed body mass into a general linear model as the dependent variable with sex and diet treatment as fixed-factor independent variables and log pronotum length as a covariate. This procedure first required testing the homogeneity of slopes assumption; if the interactions between sex and log pronotum length, and diet and log pronotum length were statistically non-significant they were removed and the ANCOVA performed.\nFor our analyses of and R i, we tested for homoscedascity among the condition scores within each treatment using Levene's test. All statistical analyses were performed in R 3.0.2 (R Development Core Team 2014) using the packages lmodel2 (Legendre 2013), smatr (Warton et al. 2011), and ggplot2 (Wickham 2009). All statistical tests were conducted at the α  = 0.05 level."}