Journal of Heredity Advance Access published online on May 19, 2008
Journal of Heredity, doi:10.1093/jhered/esn030
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
QTLs Detected in a Multigenerational Resource Chicken Population
From the Robert H. Smith Institute of Plant Sciences and Genetics, the Hebrew University of Jerusalem, Faculty of Agriculture, Food and Environmental Quality Sciences, Rehovot 76100, Israel (Atzmon, Blum, Cahaner, and Hillel); the Department of Biological Sciences, Stanford University, Stanford, CA 94305 (Feldman); and the Institute of Plant Sciences, ARO-Volcani Center, PO Box 6, Bet Dagan 50250, Israel (Lavi). Gil Atzmon is now at the Albert Einstein College of Medicine, Institute for Aging Research, Department of Medicine and Diabetes Research Center, Bronx, NY 10461 USA
Address correspondence to J. Hillel at the address above, or e-mail: hillel{at}agri.huji.ac.il.
The genetic structure of resource populations affects the power of tests to detect associations between DNA markers and complex traits. Following a chicken interline cross (White Plymouth Rock background), we produced a multigenerational resource population based on 4 pedigreed generations. In this large sibship, 265 parents have been genotyped, and their 3317 progenies have been phenotyped for BW21, BW42, breast meat weight, fat pad weight, and egg production. We developed an approach to increase test power by imposing several ways of validation including the minimization of false-positive associations. Some of our detected associations were in agreement with QTLs previously reported in the literature. A large fraction of the 81 screened markers was found to be associated with quantitative traits. We examined 729 associations, of which 150 (21%) were significant, and of these, 54 are supported by the literature. These 54 associations were identified by 42 markers (some of which are linked to each other). This finding not only supports the results obtained in our resource population but may also give some indication about their general properties.
The sequenced chicken genome and the subsequent abundance of molecular markers have led to the utilization of marker information in poultry breeding programs. However, the genetic complexity of the desired traits (e.g., growth, fat, and egg production) remains to be elucidated (Soller et al. 2006). Most economically important traits in chicken are controlled by quantitative trait genes (QTGs) located in quantitative trait loci—QTLs. Information on relevant genes or mutations included in QTG can be helpful in marker-assisted breeding programs (Abasht et al. 2006). This rationale led to the establishment of a chicken QTL database: http://www.animalgenome.org/QTLdb/chicken.html. The abundant information in this database serves to bridge between basic genetic architecture and complex traits in chicken (Soller et al. 2006).
Identification of marker–QTL linkage depends on the structure and size of the resource population as well as other factors such as genetic variation of the trait, polymorphisms of the markers, and the resolution of the genetic map. Weller et al. (1990) proposed an experimental design in which the granddaughters are phenotyped, the sires are genotyped, and the linkage between QTLs and markers is searched on the grandsires. This is referred to as the "granddaughter design"—GDD, as opposed to the "daughter design," in which, phenotyping and genotyping are done on the same animals of the daughter's generation. The main advantage of the GDD structure is in the reduction of the number of sires to be genotyped. In addition, test power and mapping resolution are expected to be improved. As a result, much genotyping effort is saved by applying GDD. The disadvantages of this approach are the need of one more generation and the need to phenotype a large number of progeny. In admixture designs, a similar strategy can be used by progeny testing of F2 or BC individuals (Soller and Beckmann 1990; Atzmon et al. 2006).
The identification of QTLs that affect economically important traits is of practical interest to breeders and of basic interest to geneticists interpreting the role and function of QTLs. Improving power by using data from a larger number of markers and high marker density throughout the genome can improve the probability of detection of true QTLs. Power can also be increased by simultaneously analyzing multiple traits or QTLs (Korol et al. 1995). Once QTLs are detected in a resource population, several steps must be taken in order to verify the significant association. Verification is achieved mainly through linkage analysis of independent resource populations (Hillel 1997).
Linkage disequilibrium between markers and the searched QTLs is a prerequisite for QTL detection and mapping, but its existence in commercial chicken lines is uncertain (Aerts et al. 2007). Although genetic variation within chicken breeds is larger than that between breeds (Wimmers et al. 2000), crosses between them to produce segregating resource populations in the next generations can result in powerful resource populations made of pedigrees over several generations. A resource population or resource family refers to a group of individuals that carry phenotypic and genetic variation for a given trait or traits and for a set of genetic markers. Such a population is used to detect markers that are linked to QTLs controlling the analyzed trait or traits. The multigenerational population (MGP) presented here, which was generated from an interline cross (Schreiweis et al. 2006), consists of 4 generations. The main advantage of this design is the production of a large number of progeny from a nucleus of variable founders, which should increase the power of tests to detect true associations between markers and complex traits. On the other hand, recombination events through the 4 generations may reduce linkage disequilibrium between markers and genes of interest and possibly reduce this power. The balance between these 2 opposing forces will be examined here.
 |
Materials and Methods |
|---|
 Top Materials and Methods ResultsDiscussionReferences |
|---|
Population Structure
Great-Grandparents' Generation
Two randomly selected females of the interline "ANAK 80" (a product of the former Israeli breeding company ANAK having a White Plymouth Rock background) were crossed to a randomly selected male of ANAK 80. These 2 genomes will be designated as A and B. In addition, one randomly selected female of line "ANAK 60" was crossed to a randomly selected male of ANAK 60 and will be designated as genome C. The progeny of these 3 crosses was used as grandparents in this pedigree.
Grandparents' Generation
A grandsire of genome A was crossed to 10 females of genome C to produce birds with genome AC. Similarly, 2 additional grandsire families have been produced: A grandsire of genome C was crossed to 8 females of genome A to produce birds with genome CA, and a grandsire of genome C was crossed to 7 females of genome B to produce birds with genome CB. Thus, 3 grandsire families and 25 granddam families were generated.
Parents' Generation
A total of 500 birds were generated from the above 3 crosses. About half of them (200 females and 37 males) served as parents in this structure.
Offspring
All 16 sires having the genome AC were crossed to 58 dams having genome CA to produce offspring of genome ACCA. In a similar manner, crosses were made between 1) CA genome sires and CB genome dams to obtain CACB genome offspring and 2) CB genome sires and AC genome dams to generate CBAC genome offspring (Figure 1). From these crosses, 3649 offspring were obtained of which 3317 were analyzed for the first recorded trait (BW21) in 3 hatches, 4 weeks apart. Carcass traits for males of the 3 hatches and "egg production" for females (hatch 2) are termed MGP1, and carcass traits of both males and females at hatch 1 are termed MGP2. Thus, males of hatch1 are found in both subgroups.
|
Genotyping
Simple sequence repeats (SSRs) were genotyped by the ABI 377 sequencer and the GENESCAN software (Perkin Elmer, Fremont, CA), using fluorescently labeled primers. The SSRs are scattered along the chicken genome. Genotyping was carried out only on the founders as follows: birds at the great grandparent (GGP) generation were analyzed for 250 markers to find polymorphic loci. Eighty-one of the polymorphic loci were chosen for genotyping 28 birds of the grandparent generation (the 3 grandsires and 25 granddams in the pedigree) and 237 individuals of the parents’ generation.
Statistical Analyses
Males of hatches 1, 2, and 3 and females of hatch 1 were slaughtered at the age of 42 days, whereas the females of hatches 2 and 3 were kept for the evaluation of egg production. The females of hatch 3 were discarded due to health problems (Mycoplasma Galisepticum) before they reached sexual maturity. Thus, data on body weight at 21 days of age were collected on females and males of all hatches. Data on body weight and on carcass traits at 42 days of age were collected on males of all hatches and on females of hatch 1. Linear model analyses were carried out on data from males of the 3 hatches (MGP1) and on data from both males and females of hatch 1 (MGP2). Thus, in MGP1, interactions between markers and hatches were feasible, whereas the interaction marker x gender was not applicable (no females in this population). In contrast, for MGP2 the opposite situation was true, interaction between markers and gender was possible, whereas the interaction marker x hatch was not applicable. Adjusted phenotypes of the progeny, free of external factors (gender and/or hatch), were obtained using the residuals from analyses in which the external factors served as main effects. These residuals that are free of external effects (except markers) were used to calculate the means of the sire families. The sire family means were used in least squares approaches as dependent variables and the markers as independent variables. The objective of these analyses was to search for associations between the phenotype of the analyzed trait and the sires’ genotypes at the marker loci. In MGP1, adjustments were made for hatches and in MGP2, adjustments were made for the gender effect. Outliers of the extreme ±2.5–3.5% were discarded from the analyses of all traits.
Each of the 81 markers was used as an independent variable in the "single marker analysis." Subsequently, the significant markers were used as a set of independent variables in multiple regression analyses, in which the stepwise procedure was applied. In this process, values of 0.25 and 0.05 were given for probabilities to enter and to leave the analysis, respectively. For missing data on the markers, we used the following approach: Half of the allele substitution effect was added to the dependent variable of the individual with the marker missing data and assigned the value "0.5" to the cell of missing data of that independent variable. For the 3 hatches of MGP1 and for the 2 genders of MGP2, we used dummy variables of 1 or 0 for presence or absence of the analyzed marker allele, respectively. Egg production data from age at first egg until the age of 258 days were collected on 288 female progeny of hatch 2 (fourth generation). Thus, the total egg production was significantly correlated with age at first egg (r = –0.54, P < 0.0001) and therefore, only the trait "total eggs" was reported here. All the statistical analyses were performed with the JMP 5.0 software (SAS Institute Inc 2002).
|
Results |
|---|
|
TopMaterials and MethodsResultsDiscussionReferences |
|---|
Marker Effects
Marker effects and significance for the 53 markers that had significant associations with any of the 5 traits are detailed in Table 1 and 2. At the first stage, we have focused only on general results, ignoring false positives due to multiple tests for the 81 markers. This aspect will be dealt with later. Twenty-five (31%) of the 81 screened markers were found to be significantly associated with BW21, 29 (36%) with BW42, 22 (27%) with breast meat weight, 15 (18.5%) with fat pad weight, and 6 (7.4%) with total eggs (Table 1). The average effect of the significant markers for BW21 is 0.14 phenotypic standard deviations (SDs), for BW42 and breast weight it is 0.19 SD, for abdominal fat weight it is 0.28 SD, and for egg production it is 0.25 SD. Values of confidence limits for the average marker effects are detailed in Table 1. When these effects were examined in pairs of traits, some markers were specific to each of the 2 traits, whereas others were common to both (Table 3). The proportion of common significant markers is high for growth traits (BW21, BW42, and breast weight). Twenty-nine markers were found to have a significant effect on BW42, 14 of them (48%) also affect breast weight, and 9 (31%) also affect BW21. Likewise, of the 27 markers associated with BW21, 11 (41%) also affect breast weight. Abdominal fat weight has 11 markers in common with BW42 (38% of the markers detected for BW42 and 69% of the markers detected for abdominal fat weight). Egg production has a small number of markers in common with the other traits (0–5).
|
|
Interactions
Markers x Hatches
The marker effects presented in Table 1 are based on analysis of males in all 3 hatches (except for the trait total eggs). Thus, it was possible to examine the interaction between markers (or QTLs) and hatches. Interactions with hatches for body weight at 21 days of age (BW21) were found for 6 markers (marked as superscripts in Table 1). Interestingly, for the other growth traits that were measured at later ages (42 days) when the effect of the hatch is expected to be reduced; only 2–3 such interactions were revealed. The trait total eggs was analyzed on females’ data of hatch 2 only.
Marker x Gender
Five interactions between markers and gender were detected; 2 markers for BW21, 3 other markers for BW42, and none for breast weight or abdominal fat weight (Table 2). In all 5 cases, the effect of the marker (difference in performance between marker alleles) was not significant among females but highly significant among males. This suggests that the analyzed QTLs function differently in males and females. Due to the data structure (i.e., females were slaughtered before maturity), this analysis was not applicable to the trait egg production.
|
Multiple Regression Analysis
If several QTGs function similarly or in the same metabolic pathways, their allele distribution among the resource progeny might be similar. As a result, in a stepwise multiple regression analysis (Tables 1 and 2), when one of the QTLs is in the regression equation, the possible contribution of the other QTLs is negligible and cannot enter the equation. We applied this procedure by using the phenotypic performance of the analyzed trait (free of hatch and gender effects) as a dependent variable and the significant markers from the single marker analysis as independent dummy variables (1 or 0). As expected, the number of significant markers was reduced considerably, the partial effect of each marker in the equation was increased up to more than 3-fold, and their statistical significance was increased accordingly. In order to take into consideration false-positive associations resulting from multiple tests of the 81 markers, we used the stringent Bonferroni correction; P value for inclusion was set to be 0.05/81 = 0.00062. Comparing these single marker analysis results (after taking into consideration the multiple tests) with the partial regression coefficients presented in Table 1 (before taking into consideration the multiple tests) shows a reduction in the total number of significant markers (P
0.0001) for the 4 traits from 19 to 10 (Table 1). The new regression equations include the most significant markers of the previous stage.
Consistency Between MGP1 and MGP2
The reported results in Table 1 (MGP1) are based on 612 males having both phenotypes and genotypes in the 3 hatches; H1 = 196 birds, H2 = 302, and H3 = 114 (Table 4). The results presented in Table 2—MGP2—are based on 427 male and female progenies having both phenotypes and genotypes of hatch 1 (Table 4). In addition to the 196 males of hatch 1 that were analyzed and presented in Tables 1 and 2, 238 females (having both phenotypes and genotypes) were analyzed and are presented only in Table 2. In spite of the overlap of 196 males of hatch 1, comparison between the 2 tables may provide some sort of support. Agreements between the 2 tables are as follows (Table 5):
|
|
- BW21—12 (48%) of the 18 associations detected in MGP2 were also detected in MGP1. Similar comparisons are given below for the other 3 broiler traits:
- BW42—4 out of 12 (33%).
- Breast weight—5 out of 15 (33%).
- Fat pad weight—3 out of 10 (30%).
- BW42—4 out of 12 (33%).
Egg production was measured on progeny of H2 only, so no counterpart was available for comparison. Inconsistencies could result either from gene fixation in 1 of the 2 resource families or as a result of false positives.
Consensus Information regarding the Associations between Markers and Traits
We further extended our analysis to identify published consensus information that is in agreement with the QTLs detected in our population. These supporting data are presented in Tables 1 and 2. One important objective of this report was to identify markers reliably associated with some production traits. Validations of the results were obtained by 1) comparison of the current results with those obtained in other resource populations, including ours (Atzmon et al. 2006), 2) comparison with results reported in the scientific literature, and 3) imposing a stringent statistical significance, P < 0.0001 (see Tables 1 and 2). Thus, markers footnoted in these tables have passed at least one of the following criteria:
- Highly significant association in one of the resource populations.
- Significant associations between markers and traits in at least 2 populations. This criterion was applied for cases in which at least one resource population is from our own study (MGP1 and MGP2), whereas the others could be either ours or from the literature. Thus, in these marked associations (i.e., as labeled in Tables 1 and 2), the proportion of false-positive associations is reduced and many of them are probably accurate.
In total, we examined 729 associations across the 2 subpopulations (MGP1 and MGP2) x 4–5 traits x 81 markers. These analyses yielded 150 (21%) significant associations among which 63 associations are validated by the above indicated criteria and are listed in Tables 1 and 2. Forty-two markers were found to be involved in these associations (some of them are linked to each other). The following pairs of markers are located within the same 10-cM regions of the chicken genetic map: markers at positions 267 and 271 cM on chromosome 1, at positions 393 and 403 on chromosome 2, at positions 4 and 9 on chromosome 3, at positions 48 and 56 on chromosome 7, and at positions 9 and 13 on chromosome 19. Of the 63 associations, 54 (86%) are supported by the scientific literature, and similarly, out of the 5 regions, 4 (80%) are also supported. This finding not only validates the results for our resource populations but also suggests that our approach can be applied to other populations as well. These 63 supported associations are distributed as follows: 42 associations with one support, 15 with 2, 5 with 3, and 1 association with 4 supports. Thus, there are 21 associations with at least 2 supports. It is interesting to note that only 3 of the verified markers (at least 2 validations) that are associated with BW21 are also associated with BW42. Furthermore, only 18 of the 43 markers for BW21 listed in Tables 1 and 2 overlap with the 41 markers of BW42. These findings are in agreement with previous evidence (Sizemore and Barbato 2002) that different sets of genes contribute to the variation in these 2 traits. Marker MCW0145 on chromosome 1, position 455, could probably be considered as a confirmed "growth QTL" as it has been validated by 4 publications as associated with BW42. In addition, 5 of the 24 markers associated with abdominal fat have been verified by at least 2 independent populations, including support from at least one other publication.
|
Discussion |
|---|
|
TopMaterials and MethodsResultsDiscussionReferences |
|---|
Associations between Markers and Phenotypes
In the GDDs (Weller et al. 1990), the marker effect was estimated in our study as a regression coefficient. In this analysis, progenies that obtained allele M1 from their sire M1/M2 were given the value 1 for the dummy independent variable, and their paternal half sibs carrying the allele M2 were given the value 0. Thus, the estimated allele effect is in fact twice the marker effect that would have been estimated as the deviation of the average of group M1 from the average of the entire population. The estimate reported in this study is similar to a breeding value in 2 respects: 1) it includes the effect of allele substitution between M1 and M2 and 2) the allele was genotyped on the sires and the trait was phenotyped on the progeny (Falconer and Mackay 1996).
Minimizing the Detection of False-Positive Associations
A major objective in linkage studies is to eliminate or minimize the number of false positive as well as false-negative detections (Hillel 1997). Higher test power can be achieved via a combination of low stringency statistical tests for the inclusion of markers that are possibly linked with QTLs controlling the analyzed trait, followed by appropriate genetic tests or stringent statistical tests for distinguishing between true and false linkage. The present study is based on single marker analysis at low statistical stringency (P < 0.1) followed by: 1) multiple regression analysis at high statistical stringency and 2) taking into account the multiple tests that were performed. We believe that the associations that remain significant through this process represent true linkage between marker loci and trait loci. The decrease in the number of markers that show significant association with traits in the multiple regression analysis is probably due to the fact that some of the QTLs function in a similar way or in the same metabolic pathways and networks. This interpretation is supported by Hasenstein and Lamont (2007), who suggested that some genes may act as a cluster although they are located on different chromosomes. Another approach used here to reduce false positives was through testing the consistency of the results between 2 partially independent subpopulations (MGP1 and MGP2). Subpopulation MGP1 was based on males of 3 hatches and MGP2 was based on males and females from hatch1 (Tables 1 and 2). Agreement between the 2 subsets varied between 20% and 44% (Table 5).
Designs and Marker Effects
The MGP presented here consists of 4 generations. In this large sibship, 3317 progenies were phenotyped, and 265 parents were genotyped. Recombination events throughout the 4 generations have presumably reduced the linkage disequilibrium between markers and genes of interest, which may in turn have reduced the test power. The large population size of this MGP contributed to the large number of associations between markers and traits, we believe that a significant fraction of these associations is true (Tables 1 and 2). Such a large number of associations detected in large resource populations has also been reported by Jennen et al. (2004). In some other reports, fewer associations have been reported, possibly due to smaller resource populations or the lack of sufficient genetic variation in the broiler traits (Van Kaam et al. 1999a, 1999b; Yonash et al. 1999; Hamoen et al. 2001; Sewalem et al. 2002; Deeb and Lamont 2003). Our findings are also supported by a simulation analysis (Atzmon et al. 2007) where we found that a moderate population size is sufficient to detect most of the QTLs affecting a quantitative trait but significant test power requires large resource populations.
It is noteworthy that analysis of body composition traits in the current large population resulted in finding 20–38% of the screened markers to be associated with at least one of the broiler traits. This is in comparison to 5–16% that were found to be associated with body composition traits in a smaller resource population based on backcrosses (Atzmon et al. 2006). Interestingly, a negative relationship was found between the size and design power of the resource population and the smallest marker effect that can be detected. In the multigenerational design presented in this study, based on 3317 phenotyped and 265 genotyped birds, as little as a 0.14–0.28 phenotypic SD effect was detectable. In contrast, in the backcross population cited above, which consists of 234 phenotypes and 84 genotypes, marker effects of 0.5–0.7 SD were needed in order to be detected (Atzmon et al. 2006). Similarly, a difference in the efficiency of the detection, with respect to other factors, was found for egg production. In a broilers x broilers resource population, 8% of the screened markers were found to be associated with egg production, whereas in the broilers x layers single-cross resource population this proportion was 32%. Apparently, this difference can be attributed to the genetic difference between the parental lines with respect to the analyzed trait (Atzmon et al. 2007).
Seven of the 12 QTLs that were detected in the current study were also detected in other independent populations reported by Atzmon et al. (2006) and in simulation results (Atzmon et al. 2007). Thus, we can conclude that the large size of the resource mapping population increases the test power to detect associations between markers and traits and decreases the proportion of false-positive detections (Atzmon et al. 2007).
Interactions
Allele x Hatches
When differences between marker alleles (representing QTL alleles) involved in growth traits among males were examined in 3 hatches, 6 and 3 interactions were found at 21 and 42 days of age, respectively (Table 1). This may be a reflection of low homeostasis and large sensitivity to different environments during the intensive growth of these broilers at early ages. In contrast, at a more stable state in higher ages, the effect of hatch on difference between marker alleles is reduced.
Allele x Genders
When differences between marker alleles were examined for their effect on growth traits among males and females of hatch 1, 5 interactions were found between gender and alleles: 2 for BW21 and 3 for BW42 (Table 2). Genetic differences in BW between males and females have been reported in several studies (Tixier-Boichard et al. 1995; Van Kaam et al. 1999a, 1999b; Deeb and Lamont 2003). The 5 allele x gender interactions reported here have a special pattern. In all cases, the merit difference between alleles is not significant among females but highly significant among males. This finding may have important genetic and breeding implications. First, it is possible that at least some of the genes controlling growth in the males and females are gender specific, and therefore, research focused on chicken growth should be carried out within genders. Second, selection for improved growth in chickens should be made within genders (Atzmon et al. 2006).
|
Footnotes |
|---|
Corresponding Editor: Jerry Dodgson
Received November 21, 2007
Accepted February 18, 2008
|
References |
|---|
|
TopMaterials and MethodsResultsDiscussionReferences |
|---|
-
Abasht B, Dekkers JC, Lamont SJ. Review of quantitative trait loci identified in the chicken. Poult Sci (2006) 85:2079–2096.
Aerts J, Megens HJ, Veenendaal T, Ovcharenko I, Crooijmans R, Gordon L, Stubbs L, Groenen M. Extent of linkage disequilibrium in chicken. Cytogenet Genome Res (2007) 117:338–345.[CrossRef][Web of Science][Medline]
Atzmon G, Blum S, Feldman M, Lavi U, Hillel J. Detection of agriculturally important QTLs in chickens and analysis of the factors affecting genotyping strategy. Cytogenet Genome Res (2007) 117:327–337.[CrossRef][Web of Science][Medline]
Atzmon G, Ronin YI, Korol A, Yonash N, Cheng H, Hillel J. QTLs associated with growth traits and abdominal fat weight and their interactions with gender and hatch in commercial meat-type chickens. Anim Genet (2006) 37:352–358.[CrossRef][Web of Science][Medline]
Carlborg O, Kerje S, Schutz K, Jacobsson L, Jensen P, Andersson L. A global search reveals epistatic interaction between QTL for early growth in the chicken. Genome Res (2003) 13:413–421.
Deeb N, Lamont SJ. Use of a novel outbred by inbred F1 cross to detect genetic markers for growth. Anim Genet (2003) 34:205–212.[CrossRef][Web of Science][Medline]
Falconer DS, Mackay TFC. Introduction to quantitative genetics (1996) Essex (UK): Addison-Wesley.
Hamoen FF, Van Kaam JB, Groenen MA, Vereijken AL, Bovenhuis H. Detection of genes on the Z-chromosome affecting growth and feathering in broilers. Poult Sci (2001) 80:527–534.
Hasenstein JR, Lamont SJ. Chicken gallinacin gene cluster associated with Salmonella response in advanced intercross line. Avian Dis (2007) 51:561–567.[CrossRef][Web of Science][Medline]
Hillel J. Map-based quantitative trait locus identification. Poult Sci (1997) 76:1115–1120.
Ikeobi CO, Woolliams JA, Morrice DR, Law A, Windsor D, Burt DW, Hocking PM. Quantitative trait loci affecting fatness in the chicken. Anim Genet (2002) 33:428–435.[CrossRef][Web of Science][Medline]
Jacobsson L, Park HB, Wahlberg P, Fredriksson R, Perez-Enciso M, Siegel PB, Andersson L. Many QTLs with minor additive effects are associated with a large difference in growth between two selection lines in chickens. Genet Res (2005) 86:115–125.[CrossRef][Web of Science][Medline]
Jennen DG, Vereijken AL, Bovenhuis H, Crooijmans RP, Veenendaal A, van der Poel JJ, Groenen MA. Detection and localization of quantitative trait loci affecting fatness in broilers. Poult Sci (2004) 83:295–301.
Korol AB, Ronin YI, Kirzhner VM. Interval mapping of quantitative trait loci employing correlated trait complexes. Genetics (1995) 140:1137–1147.[Abstract]
Lagarrigue S, Pitel F, Carre W, Abasht B, Le Roy P, Neau A, Amigues Y, Sourdioux M, Simon J, Cogburn L, et al. Mapping quantitative trait loci affecting fatness and breast muscle weight in meat-type chicken lines divergently selected on abdominal fatness. Genet Sel Evol (2006) 38:85–97.[CrossRef][Web of Science][Medline]
McElroy JP, Kim JJ, Harry DE, Brown SR, Dekkers JC, Lamont SJ. Identification of trait loci affecting white meat percentage and other growth and carcass traits in commercial broiler chickens. Poult Sci (2006) 85:593–605.
Park HB, Jacobsson L, Wahlberg P, Siegel PB, Andersson L. QTL analysis of body composition and metabolic traits in an intercross between chicken lines divergently selected for growth. Physiol Genomics (2006) 25:216–223.
Rabie TS, Crooijmans RP, Bovenhuis H, Vereijken AL, Veenendaal T, van der Poel JJ, Van Arendonk JA, Pakdel A, Groenen MA. Genetic mapping of quantitative trait loci affecting susceptibility in chicken to develop pulmonary hypertension syndrome. Anim Genet (2005) 36:468–476.[Web of Science][Medline]
SAS institute Inc. JMP statistics and graphics guide, version 5 (2002) Cary (NC): SAS Institute Inc.
Schreiweis MA, Hester PY, Settar P, Moody DE. Identification of quantitative trait loci associated with egg quality, egg production, and body weight in an F2 resource population of chickens. Anim Genet (2006) 37:106–112.[CrossRef][Web of Science][Medline]
Sewalem A, Morrice DM, Law A, Windsor D, Haley CS, Ikeobi CO, Burt DW, Hocking PM. Mapping of quantitative trait loci for body weight at three, six, and nine weeks of age in a broiler layer cross. Poult Sci (2002) 81:1775–1781.
Siwek M, Cornelissen SJ, Buitenhuis AJ, Nieuwland MG, Bovenhuis H, Crooijmans RP, Groenen MA, Parmentier HK, van der Poel JJ. Quantitative trait loci for body weight in layers differ from quantitative trait loci specific for antibody responses to sheep red blood cells. Poult Sci (2004) 83:853–859.
Sizemore FG, Barbato GF. Correlated responses in body composition to divergent selection for exponential growth rate to 14 or 42 days of age in chickens. Poult Sci (2002) 81:932–938.
Soller M, Beckmann JS. Marker-based mapping of quantitative trait loci using replicated progenies. Theor Appl Genet (1990) 80:205–208.[Web of Science]
Soller M, Weigend S, Romanov MN, Dekkers JC, Lamont SJ. Strategies to assess structural variation in the chicken genome and its associations with biodiversity and biological performance. Poult Sci (2006) 85:2061–2078.
Tixier-Boichard M, Boichard D, Groeneveld E, Bordas A. Restricted maximum likelihood estimates of genetic parameters of adult male and female Rhode Island red chickens divergently selected for residual feed consumption. Poult Sci (1995) 74:1245–1252.[Web of Science][Medline]
Tuiskula-Haavisto M, de Koning DJ, Honkatukia M, Schulman NF, Maki-Tanila A, Vilkki J. Quantitative trait loci with parent-of-origin effects in chicken. Genet Res (2004) 84:57–66.[CrossRef][Web of Science][Medline]
Van Kaam JB, Groenen MA, Bovenhuis H, Veenendaal A, Vereijken AL, Van Arendonk JA. Whole genome scan in chickens for quantitative trait loci affecting carcass traits. Poult Sci (1999a) 78:1091–1099.
Van Kaam JB, Groenen MA, Bovenhuis H, Veenendaal A, Vereijken AL, van Arendonk JA. Whole genome scan in chickens for quantitative trait loci affecting growth and feed efficiency. Poult Sci (1999b) 78:15–23.
Van Kaam JBCHM, van Arendonk JAM, Groenen MAM, Bovenhuis H, Vereijken ALJ, Crooijmans RPMA, van der Poel JJ, Veenendaal A. Whole genome scan for quantitative trait loci affecting body weight in chickens using a three generation design. Livest Prod Sci (1998) 54:133–150.[CrossRef]
Weller JI, Kashi Y, Soller M. Power of daughter and granddaughter designs for determining linkage between marker loci and quantitative trait loci in dairy cattle. J Dairy Sci (1990) 73:2525–2537.[Abstract]
Wimmers K, Ponsuksili S, Hardge T, Valle-Zarate A, Mathur PK, Horst P. Genetic distinctness of African, Asian and South American local chickens. Anim Genet (2000) 31:159–165.[CrossRef][Web of Science][Medline]
Woolliams JA, Hocking PM, Burt DW, Law A, Ikeobi CON, Morrice DR, Windsor D. Quantitative trait loci for meat yield and muscle distribution in a broiler layer cross. Livest Prod Sci (2004) 87:143–151.[CrossRef]
Yonash N, Bacon LD, Witter RL, Cheng HH. High resolution mapping and identification of new quantitative trait loci (QTL) affecting susceptibility to Marek's disease. Anim Genet (1999) 30:126–135.[CrossRef][Web of Science][Medline]
Zhou H, Deeb N, Evock-Clover CM, Ashwell CM, Lamont SJ. Genome-wide linkage analysis to identify chromosomal regions affecting phenotypic traits in the chicken. II. Body composition. Poult Sci (2006) 85:1712–1721.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||