We used mice of the Berlin Muscle Mouse (BMM) population, which had been long term selected for high body weight and high muscle mass, in order to understand the selective mechanism in livestock breeding. The founder animals of BMM were initially purchased from several pet shops in Berlin, Germany. Selection on these mice was carried out in several distinct stages. On the first stage, the mice were selected for high protein content of carcass in 23 generations at the age of 60 days. Protein content determination was done through chemical analyses. After that, they were selected in 10 generation for high body weight and low fat content at the age of 42 days. Then, mice were checked for high muscularity by palpating. The highest muscular ones on a scale of 1 to 5 were selected for the next generation. Consequently, after 25 generations, the mice were highly muscular. A high compact sub-line was perpetuated through random mating of selected animals (Varga et al. 1997). This line was sequenced for myostatin gene. It revealed that there was a 12 bp deletion which caused loss of function. (Szabo et al. 1998)
Get your grade
or your money back
using our Essay Writing Service!
After 86 generations of selection, full-sib mice with distinct phenotypes were mated. They form the basis of seven Berlin Muscle Mouse inbred lines (BMMI) four of which carry MstnCmptdl1Abc (a mysostatin mutation on MSTN gene) and three of them are wild types. We used the BMMI806 and BMMI816 lines, which are hyper-muscular but do not carry the myostatin mutation.
In the course of setting up the crossbred experiment, the mouse lines were in their 21st generation.
In order to generate F3 intercross populations, two pairs of full-sibs Berlin Muscle Mouse inbred lines BMMI806 and BMMI816 were crossed reciprocally. 94 F2 animals were randomly mated (Schmitt et al. 2009) to produce 345 F3 animals. We used 331 F3 individuals for our QTL analysis because of genotyping errors in some individuals.
Husbandry and feeding conditions
All the experimental protocols that we used for the animals were approved by the German Animal Welfare Authorities (approval no.G0405/08) and animals were treated accordingly. The animals were maintained under conventional conditions that is 22 Â± 2Â°C temperature and 12:12 hours light:dark cycles of lighting. They were caged in groups of 2 to 4 animals of the same sex per macrolon cage and had ad libitum access to food and water. The animals were fed a standard diet ('Altromin standard breeding diet no. 1314 TPF', Lage, Germany) until 70 days old. Their diet was composed of 27.0% crude protein, 5.0% crude fat, 4.5% crude fibre, 6.5% crude ash, 50.5% nitrogen free extract (starch and sugar), vitamins, trace elements and minerals (2988 kcal/kg metabolizable energy; thereof 27.0% energy from proteins, 13.0% from fat and 60.0% from carbohydrates).
The mice, at the age of 71 days, were sacrificed after a two hour fasting and being anaesthetised by isoflurane. The musculus longissimus (ML) and Musculus quadriceps (MQ) were cut up and weighed. Muscle mass (MM) was recorded as summed muscle weight of left and right M. longissimus and left and right M. quadriceps. The right muscles were immediately frozen in liquid nitrogen and then stored at -80 Â°C. Carcasses were stored at 6Â°C and pH values were taken within the M. biceps femoris at 1 and 24 hours post mortem (ebro PHT 810, Ingolstadt, Germany).
Body weight was measured weekly. We used the week ten measurement in our analyses. Fat and lean mass was measured also at week ten by quantitative magnetic resonance (QMR) analysis, using the EchoMRI whole body composition analyser (Echo Medical Systems, Houston, Texas, USA) (Neuschl et al. 2010, Tinsley et al. 2004). After the two-hour fasting, before dissection, blood glucose levels were measured. We measured the muscle glycogen content colorimetrically in the right M. longissimus (GOD/PAP method 'Glucose liquicolor' by Human, Wiesbaden, Germany) as suggested by Barham and Trinder, 1972.
Parental BMMI-lines were genotyped with the Mouse-Diversity-Array (Yang et
al. 2009) comprising 623,124 single-nucleotide polymorphisms (SNPs) at
KBiosciences (Hoddesdon, U.K.). After 21 generations of inbreeding, BMMI806 and BMMI816 lines had fixed alleles in 97.4% and 97.8% of their genomes, respectively. Both lines differed by 6.3% at the SNP level. For genotyping parents, F2 and F3 generations, according to the information on diverse genomic regions between BMMI806 and BMMI816 lines, we selected 184 markers that cover all chromosomes, except Y and X, with an average distance of 16.2 Mb (Figure 2). Regions larger than 10 Mb without informative markers did not differ in SNP-alleles between parental lines and were not included into the linkage analysis (Figure 2). The conversion of genetic map into physical map was performed via "Mouse Map Converter" software from the Jackson Laboratory (Cox et al. 2009).
Always on Time
Marked to Standard
Figure 2. Map of 164 reference single nucleotide polymorphisms used in this study. Positions are given in Mb. Bars indicate identified QTL with genome-wide significant main effect.
QTL and statistical analysis
The data analysis for this paper was conducted using SAS software version 9.1 (SAS
Institute, Cary, NC) and SPSS and PedPhase (version 2.0). PedPhase is a software package that infers the haplotypes given the genotype and pedigree of all individuals. Its algorithm is based on the minimum recombination principle. This principle says that since recombination is rare, haplotypes that cause the fewest recombinants are preferred (Gusfield 2002). Our genotypes are comprised of a combination of pairs of 1 and 2. 11 refers to A/A, 12 is for A/a and 22 is for a/a (where A shows paternal allele and a shows maternal allele). After getting the genotypes out of PedPhase, it is required to change the way they are written in order for them to be readable by SAS software. At this stage, we have index scores for additive (a), dominance (d) and imprinting (i) effect of every marker in every F2 and F3 individual (Mantey et al. 2005).
They show the estimates of the following parameters:
The genotypic values are (AA, Aa, aA, aa) and r is the reference point of the model, i.e. the midpoint between homozygotes, a is the additive genotypic value, i.e. half of the difference between two heterozygotes, d is the dominance genotypic value, i.e. the difference between the mean of heterozygotes and the midpoint of homozygotes and i is the imprinting value, i.e. the difference between heterozygotes.
To detect QTLs we performed a genome scan model with four ordered genotypes as a fixed class variable (Wolf et al. 2008). This model seeks the overall null hypothesis of no QTL. Thus, the significant QTL can have a combination of additive, dominance and imprinting (parent-of-origin effect). These effects are orthogonal in the model; hence it is possible that they are present together at a locus, without the overall effect having to be significant. QTL were characterised in terms of their pattern of effect, i.e. additive, dominance of imprinting. Therefore, this model provides unbiased description of pattern of effect of loci.
The model we used for the QTL detection was a genome scan through a linear mixed model fitted by restricted maximum likelihood (REML) with family as a random effect and the four ordered genotypes as fixed effects. In order for a more precise detection of pattern of effects, we used both original measurements and log-transformed values (Log(X)). In the interest of increasing the signal/noise ratio, we adjusted the phenotypes for sex and PGM (Parental Grandparent Mother) through univariate general linear model by SPSS. The measured phenotype value was chosen as the dependant variable and sex and PGM as the fixed variables. This adjustment significantly increased our precision in detection of QTL and patterns of effect.
Trait specific significant thresholds were calculated by 1000 permutations (Churchill and Doerge 1994). For the overall QTL effect, LOD (Log of Odds) scores higher than 3.5 were considered as significant, while for additive, dominance and imprinting effects, significant LOD scores were chosen to be higher than 1.3 (p â‰¤ 0.05) (Li et al. 2005). These scores were chosen upon the results acquired from permutation tests in which top 5% samples had a score above aforesaid thresholds. Therefore, these thresholds are specific to these samples and cannot be generalized for other studies.