| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Animal Breeding, Animal Production Research, MTT Agrifood Research Finland, Jokioinen, FIN-31600, Finland.
Animal Breeding and Genetics Group, WIAS, Wageningen University, Wageningen, The Netherlands, 6700 AH.
Department of Animal Science, University of Nebraska-Lincoln, Lincoln, NE 68583-0908.
Turku Centre for Biotechnology, P.O. Box 123, BioCity, Turku, FIN-20521, Finland.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Key Words: cofactor • dairy cattle • interval mapping • quantitative trait
Abbreviation key: BTA = Bos taurus chromosome, DYD = daughter yield deviations, FY = fat yield, F% = fat percentage, MY = milk yield, PIC = polymorphic information content, PY = protein yield, P% = protein percentage
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
In dairy cattle, QTL mapping utilizes existing half-sib breeding populations, which are routinely produced by artificial insemination. Weller et al. (1990) proposed the granddaughter design to analyze linkage between a single marker and a QTL in an outbred half-sib data structure. The idea of interval mapping originally introduced by Lander and Botstein (1989) was implemented for half-sib designs by Georges et al. (1995) and Knott et al. (1996). Several QTL mapping efforts have been undertaken in different breeds of dairy cattle by using various mapping approaches (see Mosig et al., 2001). Having QTL mapping results from cattle with different origin improves the potential to implement fine mapping strategies that take advantage of historical recombinants. The improvement of mapping resolution is an essential step towards positional cloning of mapped QTL and marker-assisted breeding schemes.
The methods described by Georges et al. (1995) and Knott et al. (1996) assume only a single QTL on a linkage group and do not take possible QTL on other chromosomes into account. Jansen (1993) and Zeng (1994) proposed methods that account for the effects of linked and unlinked QTL but are only developed for inbred line cross experiments. Recently, methods for more complex pedigrees have been suggested (Jansen, 1996; Jansen et al., 1998; Kao et al., 1999). To increase the power and the precision of QTL mapping in outbred half-sib populations de Koning et al. (2001) developed a strategy for simultaneous analysis of multiple chromosomes. In this study, we applied this method to a range of milk production traits in our granddaughter design. We also analyzed the existence of multiple QTL on the same linkage group by fitting the two QTL model to the analysis (Spelman et al., 1996; Velmala et al., 1999). In this paper, we present the results of a whole genome scan for milk production QTL in Finnish Ayrshire dairy cattle.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
For every son the EBV were obtained from the national animal model evaluation of 1998, for five milk production traits: milk yield (MY), fat yield (FY), protein yield (PY), fat percentage (F%) and protein percentage (P%). Using daughter yield deviations (DYD) instead of EBV would have been more correct, because the EBV also contain information from relatives other than the bulls’ daughters. However, the large number of daughters (from 105 to over 3000 per son) makes the difference between DYD and EBV negligible. In total, the analysis included records of ~140,000 daughters.
Genotyping and Map Construction
The sperm samples were provided by Finnish AI societies. DNA was extracted from sperm using phenol-chloroform protocol as described by Zadworny and Kuhnlein (1990).
For the genome scan, 147 microsatellite markers were selected using published bovine linkage maps (Barendse et al., 1994; Bishop et al., 1994; Ma et al., 1996; Barendse et al., 1997; Kappes et al., 1997). In addition, three candidate genes were included in the analysis. The haplotype for casein genes (
S1-, ß-, and 
-casein; BTA6) was considered as one marker (Velmala et al., 1995) and a growth hormone receptor (GHR; BTA20) 3'UTR polymorphism as one (Moisio et al., 1998). A single nucleotide polymorphism (snp) in the prolactin receptor gene (PRLR; BTA20) was also used (unpublished). The objective was to cover all 29 autosomes at ~20 cM intervals, from 2 to 14 markers per chromosome. Microsatellite amplifications were carried out using fluorescence-labeled primers. The PCR products were separated on 6% polyaclyamide gels using ALF or ALFexpress DNA sequencer (Amersham Pharmacia Biotech, Uppsala, Sweden) and actual allele sizes were determined using Fragment Manager 1.2 and ALFwin Fragment Analyzer 1.02 software (Amersham Pharmacia Biotech, Uppsala, Sweden).
Marker maps were established with maximum likelihood based programs (ANIMAP program package; Georges et al., 1995). First, the CHECKPED program was used to identify genotyping errors. Then the most likely recombination rates between adjacent markers were computed with LODTABLE. The linkage analyses were performed across families using the MAKEMAP program. This reveals the most likely order of markers. The Haldane mapping function was used to convert the recombination rates to map distances (cM).
Polymorphic information content (PIC) for each marker was calculated. Specific information about markers including PCR-conditions, primer sequences, number of alleles, PIC-values and linkage map locations are available in our web site (http://www.mtt.fi/julkaisut/cattleqtl/).
Statistical Methods
QTL mapping was performed using the multimarker regression approach (Knott et al., 1996) previously used by Spelman et al. (1996) and Vilkki et al. (1997). In short, the most likely linkage phases of the chromosomes of the grandsire were determined for every family. Then for every half-sib offspring, the probability of inheriting the alternative sire’s haplotype was calculated at fixed intervals. This conditional probability provides an independent variable on which the phenotypic values can be regressed. A QTL with an additive effect was fitted at fixed intervals (1 cM) along the linkage group by regressing the trait score on the probability. The regression analysis was nested within families and weighted with the number of daughters. For every linkage group, the presence of QTL was calculated by comparing the pooled mean squares obtained from regression within families to the residual mean square. This analysis provides F-ratios along the linkage group with the maximum value being the most likely position of QTL. For more details see Vilkki et al. (1997).
The significance thresholds and the empirical P-values were estimated with the permutation test (Churchill and Doerge, 1994). Two different significance thresholds were calculated. The chromosome-wise significant levels (Pchr) for across-family analysis and within-family analysis were obtained by 10,000 permutations. For across-family analysis genome-wise (Pgenome) significance thresholds were also established. The Pgenome were obtained as follows: Pgenome = 1 - (1 - Pchr)c, where c is the number of bovine autosomal chromosomes.
One of the benefits of the regression approach is the possibility to add cofactors to the analysis. This allows us to use the information of identified QTL to search for QTL with minor effects. The approach herein is identical to the method introduced by de Koning et al. (2001).
First, the candidate regions were identified from the initial interval mapping experiment. The QTL at 5% chromosome-wise threshold level were selected to serve as cofactors in the further analysis. The conditional probability of the QTL allele being inherited from sire to son was used as a "virtual marker." The benefit of using these virtual markers as cofactors is that any position on a linkage group can be included in the analysis in any family. Secondly, multiple linear regression was used to reestimate the combined effects of all cofactors. Third, the phenotypic data was adjusted for the effects of cofactors, separately for each linkage group, using unlinked cofactors. All linkage groups were reanalyzed with the adjusted phenotypes. Chromosome-wise significance thresholds were determined by permutations (Churchill and Doerge, 1994; Doerge and Churchill, 1996). When new significant candidate regions were identified a new set of cofactors was selected and interval mapping performed. The analysis was repeated until no new candidate regions were detected and the estimated locations of QTL became stable.
The existence of multiple QTL on the same linkage group was tested by fitting a two QTL model to the analysis (Spelman et al., 1996; Velmala et al., 1999). First, test statistics were calculated for one QTL vs. none, then for two QTL vs. none. The empirical thresholds were determined with permutation test (Churchill and Doerge, 1994; Doerge and Churchill, 1996) as previously. If the test statistics for two QTL vs. none were significant, an F-test for two QTL vs. one QTL was applied. This allows us to define whether the two QTL explain more variation than one QTL. The significance of the test statistics was determined by a standard F-table. In our data, the existence of two QTL were analyzed while fitting cofactors on other linkage groups.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
). About 92% of the bovine genome was covered at ~20 cM intervals.
|
). Most of the chromosomes were relatively informative with the range between markers being 0.37 to 0.68. The average information content at genome level was 0.53.
In the across-family analysis without cofactors 2 QTL exceeding 5% genome-wise significance thresholds and 14 QTL exceeding 5% chromosome-wise thresholds were identified (Table 2). The most striking results were found on Bos taurus chromosome (BTA) 14. A genome-wise significant QTL (Pgenome < 0.0029) for F% was detected at the centromeric end of the chromosome. At the same position a QTL for FY (Pchr = 0.0050) was detected. The analysis within families suggests that the inheritance of ILSTS039 allele 14 (242 bp) correlates with an increase of F% in three families and of FY in one family. In family 9, an exceptionally high test statistic of 22.62 was observed for fat percentage (Figure 1). In this family, the size of the QTL effect for F% was 0.65%, which is almost 2.5 standard deviations (F% standard deviation for the unselected AI bull population in 1998 data is 0.27%). The QTL effect for F% in family 4 was 0.36% and in family 5 0.29%. In family 9, the effect for FY was 21 kg, which approximates more than one standard deviation. An effect for P% was also detected in family 9 and in family 6, but the position of the QTL was different. QTL affecting different milk constituents in BTA14 have been previously reported in several studies (Ashwell et al., 1998a; Ashwell et al., 1998b; Coppieters et al., 1998; Heyen et al., 1998; Ron et al., 1998; Heyen et al., 1999; Riquet et al., 1999; Moisio et al., 2000).
|
|
The effect of adding cofactors to interval mapping can be seen in Table 3. In this study the selection criteria for a QTL to be included as a cofactor was 5% chromosome-wise significance. The analysis of individual chromosomes revealed 14 such suggestive QTL (Pchr < 0.05; Table 2). When cofactors were used, the number of identified QTL for milk yield increased from 5 suggestive to 12 genome-wise significant QTL (Pgenome < 0.05). Similarly, four new regions were identified for F% and for P% and two for PY. For FY one additional QTL was identified at BTA18 but the effect was only significant at 5% chromosome-wise level. De Koning et al. (2001) also analyzed milk yield and, using cofactors, went from 5 suggestive to 8 significant QTL (BTA1, 5, 6, 12, 20, 21, 23, 29), using virtually the same data. In addition to the QTL reported by de Koning et al. (2001) we found four other significant QTL (Table 3). The difference between these two studies is that de Koning et al. (2001) analyzed 28 chromosomes instead of 29 because genotypes for BTA3 were not available at that time. In this study, BTA3 reveals a significant QTL for MY and is thus selected to serve as a cofactor.
|
and 3 shows that the power of the analysis increases when mapping is augmented by cofactors. The test statistics increase because the residual variance decreases. The position of QTL seems to become more accurately estimated. The biggest change in test statistics was seen in the F% QTL at BTA14 where the F-ratio increased from 5.83 to 10.74 and the position remained the same. In BTA20 the test statistic for MY QTL was also noticeably increased but in this case the position was slightly changed. Figure 2 illustrates that addition of cofactors not only increases the test statistics and changes the best position of the QTL but also makes the F-ratio peak steeper.
|
, it is quite common that a QTL detected in one milk production trait reappears for another trait. Some caution should be taken before drawing any final conclusions about the detected milk production QTL, especially F% and P%, without fully understanding the true nature of the trait. Regarding chromosome areas with QTL for several milk traits at the same position, the effect on MY and F% or P% (or both) are usually opposite (data not shown). This means that either the high QTL allele for MY is on the same chromosome phase with the low F% or P% (or both) alleles and vice versa, or it reflects the "pleiotropic" effect of one gene. We assume that the percentage traits are most often reflecting the amount of water in milk as the increase in milk water content decreases the proportion of milk solids. We would like to point out, that depending on the power of the analysis, in several cases the true nature of the effect could be easily missed by missing one of the correlated effects. The cofactor analysis increases the power of detection of QTL noticeably. Interestingly, in our data the effect seen in MY and F% or P% (or both) is never reflected in fat and protein yields except in BTA25, where QTL for MY, P% and PY were detected.
The closer study of within-family results supports the idea that many of the QTL affect the liquid component of milk. The tendency can be seen in chromosomes 1, 3, 6, 20, 21, and 23. Similar observations have been made in the Holstein-Friesian population for chromosomes 6, 9, 14, 20, and 26 (Zhang et al., 1998). Because lactose is the major osmotic molecule in milk it is tempting to suggest that many QTL detected for milk yield are due to the genetic factors related to lactose synthesis and secretion. However, it should not be forgotten that lactose is not the only osmotic component of milk. Some minerals, especially calcium and phosphorus, affect the osmotic potential, too. They are usually complexed with the main protein components of milk, casein and casein micelles.
On BTA12 and BTA14, genome-wise significant QTL clearly affecting yield were found. The QTL effects for MY, FY, and PY in BTA12 are parallel. The single chromosome analysis of BTA12 revealed QTL for FY in two families (3 and 7) and also QTL for MY and PY in family 7. After cofactor analysis the latter two were also seen in family 3 (Pchr < 0.05). The allele substitution effects in this family for MY and PY was, respectively, 192 and 5.2 kg. This is the only QTL detected in Finnish Ayrshires that has an influence on the overall yield. The cofactor analysis did not reveal any additional information at family level for fat traits in BTA14.
The distribution of detected QTL among different chromosomes and families (Figure 3) shows a rather disappointing picture considering the practical applications of the results. First, the assumption that the majority of the QTL affect the liquid component of milk seems to be supported by the large number milk yield and percentage QTL compared with fat and protein yield QTL. Second, for the genuine yield QTL, only few families share each QTL, hampering both further fine mapping and use of marker information in breeding.
|
Several QTL were detected in family 9, some with very high test statistics. Some caution should be taken regarding results from such families, particularly when the family size is small. In this family, the allele substitution effect in BTA14 is similar to other segregating families and thus the QTL seems to be genuine. However, in some chromosomes the observed QTL effect may be due to having information only on few individuals.
We also analyzed the possible presence of two QTL on the same linkage group, while correcting for QTL on other linkage groups. Suggestive evidence for the presence of two QTL was found only on chromosome 3 (F = 2.08). One QTL was located at the centromeric end of the chromosome (1 cM) and the other at the distal end of the chromosome (106 cM). In our previous study, two QTL were detected in BTA6 for P% and MY (Pchr < 0.1) and for F% (Pchr < 0.05; Velmala et al., 1999). In this study no significant support for our previous findings was observed. One difference between these two studies is that Velmala et al. (1999) used EBV from the national animal model evaluation of 1996, whereas the EBV in this study are from the evaluation of 1998. Another difference between these two studies is that here the existence of two QTL was jointly analyzed with the cofactors. Possibly, the earlier result was due to linkage disequlibrium extending over different chromosomes, creating spurious association, which was removed by cofactors.
In Finnish Ayrshires, several loci affecting milk production traits are still segregating. Most of the QTL studies in dairy cattle have been carried out in different Holstein-Friesian populations (see Mosig et al., 2001). Interestingly, the analysis of the red and white cattle breed reveals only one QTL not detected in Holstein-Friesians or other dairy populations. Recent studies on genetic relationships among cattle breeds have suggested a common ancestry for European Ayrshire and Friesian breeds (Kantanen et al., 2000). An interesting question is whether the QTL detected so far have the same origin in different breeds of cattle. However, it is difficult to compare studies to acquire comprehensive results because different studies have used different designs, methods, traits, and levels of statistical significance. Moreover, the physiological nature of milk production traits makes comparative analysis even more complicated because some of the studies cover only one or two traits.
Drawing conclusions about the detected QTL should be done with caution because some effects may have remained undetected. This is probably due to the relatively low power of analyses used for QTL detection so far. The augmentation of cofactors not only increases the power to detect QTL but also clarifies the nature of milk production QTL found. A further improvement of mapping precision and power would be gained by taking into account effects of the putative QTL on several traits simultaneously (Korol et al., 2001).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Corresponding author: S. M. Viitala; e-mail:
sirja.viitala{at}mtt.fi.
Received for publication July 13, 2001. Accepted for publication January 30, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Momke, H. Scholz, K. Doll, J. Rehage, and O. Distl Mapping Quantitative Trait Loci for Left-Sided Displacement of the Abomasum in German Holstein Dairy Cows J Dairy Sci, November 1, 2008; 91(11): 4383 - 4392. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Abe, J. Saburi, H. Hasebe, T. Nakagawa, T. Kawamura, K. Saito, T. Nade, S. Misumi, T. Okumura, K. Kuchida, et al. Bovine quantitative trait loci analysis for growth, carcass, and meat quality traits in an F2 population from a cross between Japanese Black and Limousin J Anim Sci, November 1, 2008; 86(11): 2821 - 2832. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. P. P. Macciotta, M. Mele, G. Conte, A. Serra, M. Cassandro, R. Dal Zotto, A. Cappio Borlino, G. Pagnacco, and P. Secchiari Association Between a Polymorphism at the Stearoyl CoA Desaturase Locus and Milk Production Traits in Italian Holsteins J Dairy Sci, August 1, 2008; 91(8): 3184 - 3189. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. D. Daetwyler, F. S. Schenkel, M. Sargolzaei, and J. A. B. Robinson A Genome Scan to Detect Quantitative Trait Loci for Economically Important Traits in Holstein Cattle Using Two Methods and a Dense Single Nucleotide Polymorphism Map J Dairy Sci, August 1, 2008; 91(8): 3225 - 3236. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Lipkin, A. Bagnato, and M. Soller Expected Effects on Protein Yield of Marker-Assisted Selection at Quantitative Trait Loci Affecting Milk Yield and Milk Protein Percentage J Dairy Sci, July 1, 2008; 91(7): 2857 - 2863. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Lipkin, R. Tal-Stein, A. Friedmann, and M. Soller Effect of Quantitative Trait Loci for Milk Protein Percentage on Milk Protein Yield and Milk Yield in Israeli Holstein Dairy Cattle J Dairy Sci, April 1, 2008; 91(4): 1614 - 1627. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Viitala, J. Szyda, S. Blott, N. Schulman, M. Lidauer, A. Maki-Tanila, M. Georges, and J. Vilkki The Role of the Bovine Growth Hormone Receptor and Prolactin Receptor Genes in Milk, Fat and Protein Production in Finnish Ayrshire Dairy Cattle Genetics, August 1, 2006; 173(4): 2151 - 2164. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Khatib, S. D. Leonard, V. Schutzkus, W. Luo, and Y. M. Chang Association of the OLR1 Gene with Milk Composition in Holstein Dairy Cattle J Dairy Sci, May 1, 2006; 89(5): 1753 - 1760. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Y. Chen, Q. Zhang, C. C. Yin, C. K. Wang, W. J. Gong, and G. Mei Detection of Quantitative Trait Loci Affecting Milk Production Traits on Bovine Chromosome 6 in a Chinese Holstein Population by the Daughter Design J Dairy Sci, February 1, 2006; 89(2): 782 - 790. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Szyda, Z. Liu, F. Reinhardt, and R. Reents Estimation of Quantitative Trait Loci Parameters for Milk Production Traits in German Holstein Dairy Cattle Population J Dairy Sci, January 1, 2005; 88(1): 356 - 367. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Boettcher, A. Caroli, A. Stella, S. Chessa, E. Budelli, F. Canavesi, S. Ghiroldi, and G. Pagnacco Effects of Casein Haplotypes on Milk Production Traits in Italian Holstein and Brown Swiss Cattle J Dairy Sci, December 1, 2004; 87(12): 4311 - 4317. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. F. Schulman, S. M. Viitala, D. J. de Koning, J. Virta, A. Maki-Tanila, and J. H. Vilkki Quantitative Trait Loci for Health Traits in Finnish Ayrshire Cattle J Dairy Sci, February 1, 2004; 87(2): 443 - 449. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
