Detection of Quantitative Trait Loci Affecting Milk Production Traits on ...

25 downloads 75 Views 137KB Size Report
(Georges et al., 1995; Lipkin et al., 1998; Spelman et al., 1996; Velmala et al., ...... G. A. Hawkins, S. S. Toldo, R. Fries, M. D. Grosz, and J. Yoo. 1994. A genetic ...
J. Dairy Sci. 84:1516–1524  American Dairy Science Association, 2001.

Detection of Quantitative Trait Loci Affecting Milk Production Traits on 10 Chromosomes in Holstein Cattle Y. Plante,†,1 J. P. Gibson,*,1,2 J. Nadesalingam,* H. Mehrabani-Yeganeh,* S. Lefebvre,* G. Vandervoort,* and G. B. Jansen* *Centre for Genetic Improvement of Livestock Animal and Poultry Science, University of Guelph, Guelph, Ontario N1G 2W1 †Bova-Can Laboratories, Saskatchewan Research Council, 15 Innovation Blvd., Saskatoon, Saskatchewan S7N 2X8

ABSTRACT Sons (n = 71 to 75) of each of six Holstein sires were genotyped at 69 microsatellite loci covering a total of 676 cM on chromosomes 3, 5, 9, 10, 13, 15, 17, 20, 23, and 26. Estimates of quantitative trait loci (QTL) effect and location were made using a least squares interval mapping approach based on daughter yield deviations of sons for 305 d milk, fat, and protein yield and fat and protein percentage. Thresholds for statistical significance of QTL effects were determined from interval mapping of 10,000 random permutations of the data across the bull sire families and within each sire family separately. Analyses combining data across sires indicated the presence of QTL affecting milk, fat, and protein yield on chromosomes 20 and 26 and a QTL affecting fat and protein percentage on chromosome 3. Analyses within each sire family separately indicated the presence of segregating QTL in at least one family on 7 of the 10 chromosomes. Statistically significant estimates of QTL effects on breeding value ranged from 438 to 658 kg of milk, from 17.4 to 24.9 kg of fat, 13.0 to 17.0 kg of protein, 0.04 to 0.17% fat, and 0.07 to 0.10% protein. (Key words: quantitative trait loci, marker maps, milk production, cattle) Abbreviation key: DYD = daughter yield deviations, LS = least squares, MARC = Meat Animal Research Centre. INTRODUCTION

velopment of reasonably dense microsatellite linkage maps for the bovine genome (Barendse et al., 1997; Bishop et al., 1994) have made marker mapping of QTL a practical reality. Several studies recently have reported the presence of significant QTL affecting milk production traits on several different chromosomes (Georges et al., 1995; Lipkin et al., 1998; Spelman et al., 1996; Velmala et al., 1999; Zhang et al., 1998). The sizes of QTL effects being discovered are more than sufficient to warrant their use in selection programs, particularly for the preselection of young bulls entering progeny testing (Gomez-Raya and Gibson, 1993; Kashi et al., 1990). We report here on the use of microsatellite markers and a granddaughter design (Weller et al., 1990) to map QTL contributing to variation in milk production traits on 10 chromosomes in six Holstein sire families. The results confirm the existence of several QTL detected in previous studies as well as detecting several QTL previously unreported. MATERIALS AND METHODS Overview A total of 434 sons of six prominent Holstein sires, with 71 to 75 sons per sire family, were genotyped at two to nine microsatellite loci on chromosomes 3, 5, 9, 10, 13, 15, 17, 20, 23, and 26. Least squares interval mapping was performed based on daughter yield deviations (DYD) for 305-d milk, fat and protein yield, and fat percent, and protein percent.

The discovery of highly polymorphic microsatellite markers (Litt and Luty, 1989) and the subsequent de-

DNA Preparation and Genotyping Methods

Received February 24, 2000. Accepted December 31, 2000. Corresponding author: J. P. Gibson; e-mail: jgibson@aps. uoguelph.ca. 1 These authors contributed equally to this paper. 2 Current address: ILRI, P.O. Box 30709 Nairobi, Kenya.

Minor modifications of standard methods were used for DNA extraction and storage and for genotype assays of microsatellites using radioactive end-labeled primers as fully described in Nadesalingam et al. (2000). All gels were scored independently by three people, and inconsistencies between scorers were checked against

1516

1517

QTL FOR PRODUCTION TRAITS OF HOLSTEIN CATTLE

the photograph of the original gel. Where genotypes remained unclear or were not possible given the sire genotype, the assay was repeated or the genotype was defined as unknown. Where allele frequencies within sire families approached significant deviation from expected given the sire’s genotype and the estimated allele frequencies in the dam population, the original gels were scored again and where results seemed open to error, the whole family was assayed again for the genotype in question. Performance Data All QTL mapping analyses were based on DYD (VanRaden and Wiggans, 1991) of sons based on progeny tests that generally involved from 50 to 100 daughters. DYD were supplied by the Canadian Dairy Network based on Canadian calculations of EBV in August, 1997. DYD for fat percent were derived from DYD for milk and fat yield by adding the base population mean for milk and fat yield to each DYD, and then dividing the fat DYD by the milk DYD. The DYD for protein percent was derived similarly. Linkage Mapping Linkage orders and map distances among markers were estimated using the CRI-MAP program (P. Green, 1990, personal communication) with map distances based on Kosambi’s mapping function. Results were compared to the Meat Animal Research Centre (MARC) linkage map as presented on their web site (http://sol.marc.usda.gov/genome/cattle/cattle.html). When our order differed from the MARC linkage map order, the correct order was deemed to be the one on the MARC map (see below for justification). We then estimated recombination rates and distances for the order deemed correct. Double and triple crossover events were identified to locate possible genotyping errors. When unlikely sets of multiple crossover events occurred involving a particular marker, the genotypes were either scored again or the genotypes were set as unknown and the map order and distances were reestimated. Only four chromosomes’ genotypes were edited at this stage, with no more than five genotypes altered for a given chromosome. Statistical Analyses A weighted least squares (LS) interval mapping was performed using a modified version of the program developed by S. Knott and C. Haley, with details as described in Knott et al. (1996). The combined-sire model for the analysis was,

DYDij = sj + bj pij + eij, where sj is a fixed effect for sire j, bj is the regression coefficient for the QTL nested within sire j, pij is the probability of inheriting the defined QTL allele from sire j for son i at a given position, and eij is a residual error, with variance approximately equal to σe2/Rij, where Rij is the reliability of the DYD of son i. The weighting factor in the weighted least squares analysis was Rij. Analyses were performed at 1-cM intervals along the chromosome. Separate analyses were performed for 305-d milk, fat and protein yield, and fat and protein percentage. Interval mapping was also performed for each of the five traits within each sire family separately (referred to here as within-sire analyses). For all analyses, the map distances between markers were taken from the current MARC map at the time of the analysis. The justifications for doing this are, 1) the MARC map, with its much higher marker density and genotypes of both parents, should have a much lower genotyping error rate and should therefore be more accurate, 2) by doing so all map positions of QTL are expressed on an internationally recognized and easily accessible map, and 3) it was previously demonstrated that it made essentially no difference to interpretation of results whether we used our map distances or those of MARC (Nadesalingham, 1999). Thresholds for testing significance of effects were obtained by 10,000 random permutations of the DYD data within sire families, with a full combined-sire and within-sire analysis repeated for each trait × chromosome combination for each permutation, as suggested by Doerge and Churchill (1996). The highest F ratio for each analysis was stored and ranked, and the 1, 5, and 10% thresholds were found as the 100th, 500th, and 1000th ranked F ratios for each chromosome by sire (or combined-sire) by trait combination. RESULTS The markers used in these analyses are shown in Table 1. The numbers of markers that were informative for each sire by chromosome combination are shown in Table 2. Table 1 also shows the estimated map distances between markers based on the present study and on the MARC linkage map at the time of the analyses. The most likely linkage order from the present data differed from that of the MARC map only for chromosomes 5 and 15, and in both cases the MARC order had a LOD score less than 0.1 higher than the most likely order derived from our data. The MARC maps for chromosomes 3 and 10 did not include all the markers used in the present study. For these two chromosomes, Table 1 presents “consensus” Journal of Dairy Science Vol. 84, No. 6, 2001

1518

Journal of Dairy Science Vol. 84, No. 6, 2001

Table 1. Marker loci and map distances1 for 10 chromosomes. Chromosome 3 5 9 10 13

17 20 23 26

INRA006 19.5 0.0 BM6026 6.7 0.0 ETH225 8.1 0.0 CSSM38 7.0 0.0 BMC1222 19.5 0.0 ACDY2 20.7 0.0 ETH185 62.3 0.0 BM3517 0.0 0.0 BM47 9.1 0.0 HEL11 20.7 0.0

FCGR1 15.5 5.8 BP1 12.1 7.7 BM1227 11.0 22.4 BM1237 12.6 12.9 BMS1145 15.7 24.8 HEL1 7.0 15.3 BM1233 36.3 80.0 BM1225 8.0 9.3 BM1258 14.8 30.3 BM4505 19.0 23.9

INRA023 19.2 14.7 BMC1009 21.8 31.7 BM2504 6.1 2.8 BM6305 16.6 23.3 BL42 23.7 52.0 HBB1 11.9 26.4

TGLA126 23.2 14.5 CYP21 12.1 14.0 BM188 0.7 0.0

INRA003 5.3 13.5 MAF23 22.9 26.1 ILST013 18.5 21.6 TGLA4 2.3 10.8

BR4502 43.7 58.0 IGF-1 10.5 14.4 BMC701 13.0 19.1 BM888 11.9 29.6

BM2924 13.3 18.8 BM1819 3.6 1.8 BM6436 14.7 12.6 TGLA102 2.5 3.3

BM315 22.5 15.8 BM4208 13.2 11.2 CSRM60 20.2 19.2

ETH152 18.2 31.4

TGLA272 13.7 18.9

CSSM39 6.1 4.5

BM848 42.0 42.2

BM713 3.1 20.6 BM1818 14.9 17.0 BM804 19.2 23.9

BM4107 18.1 6.2 BM1905 13.4 27.5

BM5004 11.9 26.5

Total MARC map Total present map 1

Inflation (%)

97.0 110.8

14.2

111.6 128.9

15.5

76.5 89.7

17.3

85.9 122.5

42.6

39.4 76.8

94.9

60.9 83.9

37.8

36.3 80.8

122.6

64.3 77.1

19.9

65.2 88.8

36.2

38.9 47.8 676.0 907.1

22.9 34.2

Map distances are sequential distances between loci in centimorgans, except for the first locus on the Meat Animal Research Center (MARC) map, which is the position relative to the most distal marker on that map. All distances are based on Kosambi’s mapping function. The MARC map was that available on the MARC web site at time of analysis during May to August, 1999. Consensus map distances are MARC distances where available, with other distances, shown in bold italics, equal to the IBRP97 distances divided by the degree of inflation of the IBRP97 map compared to the MARC map within the interval of the nearest flanking markers on the MARC map. 2 Distance in centimorgans between the external markers included here. The total MARC map contains many additional markers and is generally longer than shown here.

PLANTE ET AL.

15

Locus Consensus Present Locus MARC Present Locus MARC Present Locus Consensus Present Locus MARC Present Locus MARC Present Locus MARC Present Locus MARC Present Locus MARC Present Locus MARC Present

Total map length2

Locus and map distances

1519

QTL FOR PRODUCTION TRAITS OF HOLSTEIN CATTLE

map distances, where gaps in the MARC map were replaced by distances obtained from the IBRP97 map (http://www.ri.bbsrc.ac.uk/cgi-bin/mapviewer?species =cattle), scaled proportionally to fit the distance between the nearest flanking markers on the MARC map. The total map length estimated from our data is inflated by a total of 34.2% (the range for individual chromsomes is 14 to 123%) compared with the MARC map. This degree of inflation seems comparable to that of similar maps available on various web sites and in the published literature, and probably reflects the greater difficulty in detecting genotype errors in sparse maps, especially when only one parent is genotyped. Figure 1 shows the F values for combined-sire analyses of the 10 chromosomes. Locations are measured in centimorgans from the first marker (see Table 1) based on MARC map distances. Thresholds of F values for significance at the 5% level for combined-sire analyses ranged from 2.4 to 2.9, and at the 1% level ranged from 3.2 to 3.6. Exact probabilities do not change rapidly with F values, so it can safely be assumed that the 5 and 1% probability thresholds occur at about 2.6 and 3.4, respectively. On chromosome 3, we found evidence for a QTL affecting protein percent (P = 0.015), fat percent (P = 0.042), and milk yield (P = 0.052) located at 27, 34, and 39 cM, respectively. On chromosome 20, there was evidence for a QTL affecting protein yield (P = 0.012) at 19 cM with an effect on milk yield that approached significance (P = 0.076) at 21 cM. On chromosome 26, there was evidence for a QTL affecting fat yield (P = 0.011) at 38 cM, with a correlated effect on protein yield at the same location, which approached significance (P = 0.10). The analyses of individual sire families provide considerably greater detail on location and effects of QTL and point to the presence of QTL that are not detected by the combined-sire analyses. For brevity, we have not included figures for the individual sire analyses, but these are available at http://cgil.uoguelph.ca/pub/jdsplante.htm. Table 3 gives estimates of QTL effects on DYD at all peak F values that achieved a corresponding

probability 0.1 or less for at least one trait. The exact significance levels of the significant trait at each position are also given in Table 3, based on the corresponding permutation tests for each trait × chromosome × sire combination. The choice of P < 0.1 for inclusion in Table 3 allows a search for consistent effects across traits and sire families, albeit with an expected increase in the proportion of putative QTL that are false positives (type I errors). There are 18 significant (P < 0.05) trait × sire × chromosome effects, four of which are significant at P < 0.01. Based on the approach to statistical significance in two families (P < 0.1) and consistent estimates of QTL effects and location, it seems likely that the QTL on chromosome 3 is segregating in three of the sire families. This QTL seems primarily to affect milk yield, with corresponding effects on fat and protein percentage. Similarly, the QTL on chromosome 23 appears to be segregating in three of the sire families, and seems to have a general effect on milk, fat, and protein yield. The QTL on chromosome 26 seems most likely to be the same QTL in at least two sire families. This QTL has a general effect on milk fat and protein production, but with a more pronounced effect on fat yield such that fat percentage also changes. Among the chromosomes with no significant QTL detected in the combined-sire analyses, a QTL segregating on chromosome 20 in families 1 and 2 is clearly evident. Located somewhere in the region from 10 to 40 cM, this QTL has consistent effects on milk, fat, and protein yield in the two families. The putative QTL that approached significance in family 3 seems to have rather different effects and is located about 30 cM away from the QTL in families 1 and 2. This suggests that the putative QTL in family 3 is either a false positive or may be a different QTL to the QTL in families 1 and 2, although the possibility that it is the same QTL as in families 1 and 2 cannot be completely ruled out. On chromosome 5, a QTL is evident at about 0 cM, which primarily affects protein yield and protein percentage, and is segregating in families 2 and 3; but the

Table 2. Number of informative markers by sire. Chromosome Sire

1

3

5

6

9

1 2 3 4 5 6 Average Maximum

7 6 3 6 5 3 5.0 8

5 5 4 5 3 4 4.3 6

4 5 6 4 4 5 4.7 8

5 5 2 5 5 3 4.2 7

6 3 3 4 3 4 3.8 7

10 6 6 3 6 3 6 5.0 9

13 2 1 2 1 3 3 2.0 3

15 4 3 2 2 2 2 2.5 4

17 2 1 2 0 2 1 1.3 2

20 4 3 5 3 4 3 3.7 6

23 4 4 2 3 3 5 3.5 5

26 2 4 3 1 1 4 2.5 4

Total 51 46 37 40 38 43 42.5 69

Journal of Dairy Science Vol. 84, No. 6, 2001

1520

PLANTE ET AL.

Figure 1. The F ratios for five traits across 10 chromosomes from combined-sire analyses. Milk yield, —䊏—; fat yield, —䊉—; protein yield, —▲—; fat %, —䊊—; and protein %, —✶—.

Journal of Dairy Science Vol. 84, No. 6, 2001

1521

QTL FOR PRODUCTION TRAITS OF HOLSTEIN CATTLE Table 3. Estimates of QTL effects at position of maximum F ratio for each trait and sire combination. Estimated effect on DYD2 Chromosome 3

Location1

Sire

Milk

34 35

1 1

−219 −219+

Fat

Protein

Fat %

Protein %

Exact P

−2.3 −2.3 2.0 −5.7 3.1

0.134** 0.133 −0.097 0.069 0.039

0.050** 0.049 −0.043+ 0.023 0.043*

0.004, 0.005 0.082 0.053 0.058 0.042

1

4

197

40 0

6 2

−247+ −23

4.4 4.3 −2.2 −2.3 2.4

0 63

3 3

106 282

5.9 12.4*

6.5+ 7.4

0.020 0.020

0.031 −0.016

0.07 0.026

9

44 60

2 2

287 300*

6.2 6.9

4.6 5.4

−0.048 −0.046

−0.048+ −0.045

0.063 0.027

10

85 62 28 28 33 0 28 33

5 6 5 1 1 2 1 1

−54 −329* −200 262 241+ 87 −266 −256+

8 16 64

2 2 3

182 213 40

5.8 −3.2 −10.5+ 4.7 4.3 1.6 −6.1 −5.8 8.7* 9.9 6.9

1 2 3 2 3 3 3

90 220+ 218 103 −189 −252 −245*

10.2+ 6.3 10.3* 11.9** −10.4** −10.0 −8.3

0.076 0.091 −0.029 −0.054 −0.049 −0.018* 0.041 0.041 0.024 0.026 0.055 0.074 −0.018 0.024 0.086* −0.036 −0.008 0.008

0.040* 0.036 0.029 −0.002 −0.004 0.009 0.016 0.017 0.013 0.021 0.034+ 0.017 −0.029 −0.009 0.021 −0.003 0.005 0.009

0.015 0.050 0.069 0.043 0.054 0.023 0.043 0.054 0.045 0.016

28 6 47 38 3 11 15

2.2 −7.1 −3.6 8.2* 7.3 3.6 −7.0* −6.5 6.9 8.5* 4.7 4.6 4.2 6.1 5.2 −6.3 −7.5* −6.9

5

13 17

20

23

26

0.079 0.094 0.091 0.039 0.004, 0.048 0.01 0.032 0.042

= P < 0.10; * = P < 0.05; ** = P < 0.01. Location is distance (cM) from the first marker listed in Table 1. 2 Daughter yield deviations. + 1

evidence for this QTL is weaker than that for chromosome 20. Also on chromosome 5, another QTL is evident in family 3 at 63 cM, which seems to affect milk, fat, and protein production. On chromosome 10 there is evidence for a single QTL at about 60 to 80 cM, with variable effects on milk versus fat and protein yield such that fat and protein percentage are affected. This QTL appears to be segregating in families 5 and 6. On chromosome 17 it is possible that the QTL detected in families 1 and 2 is the same QTL, probably affecting milk and protein yield with a lesser effect on fat yield such that fat percentage is affected. The QTL detected on chromosome 9 at about 40 to 60 cM in family 2 appears to affect milk yield more than fat and protein yield such that fat and protein percentage are also affected. Although results did not approach statistical significance, sire 1 also showed peaks in the F ratios for milk yield, fat, and protein percentage at about the same location as the significant QTL in family 2 (see figures at http://cgil.uoguelph.ca/

pub/jds-plante.htm). The QTL on chromosome 13 in family 5 did not achieve statistical significance (P = 0.069) and there was no suggestion of a similar QTL segregating in other families. DISCUSSION Based on our interpretation of which QTL identified in Table 3 likely represent the same QTL in different families, and excluding the weak evidence of chromosome 9, we have interpreted the results as indicating a total 13 segregating QTL × sire combinations, corresponding to a total of eight QTL on seven chromosomes. A critical question in QTL mapping studies is how many of the statistically significant effects represent real QTL rather than false positive results. Allowing for the number of trait × sire × chromosome combinations that were examined, one would expect a total of 5 × 6 × 10 × 0.05 = 15 significant estimates at P < 0.05, compared with the 18 observed. The fact that Journal of Dairy Science Vol. 84, No. 6, 2001

1522

PLANTE ET AL.

the traits are correlated does not affect the above expectation, but it does impact the expectation of the number of sire × chromosome combinations for which a QTL is inferred based on the statistical significance of one or more traits at similar locations. A principal components analysis of the covariances among the DYD of the five traits for the 440 bulls used here revealed that the first three principal components explained 48.6, 39.0, and 10.5% of the variance, respectively. Thus, information content of this data is equivalent to that of a little more than two independent traits. For two independent traits, the probability that a significant QTL is inferred, based on one or both trait effects being significant at a consistent location, for a given sire on a given chromosome is a little higher than 1 − 0.952 = 0.097. Given that the present data is equivalent to a little more than two independent traits, the number of QTL × sire combinations expected to be inferred for the present experiment is a little higher than 6 × 10 × 0.097 = 5.8, which compares to the 13 observed. This line of reasoning indicates that the number of inferred QTL × sire combinations based on significance of one or more traits was a little less than 2.2 times higher than expected by random chance. More compelling evidence for the effects being real QTL comes from the detection of QTL with similar effects in similar locations in more than one sire family on seven of the 10 chromosomes. Indeed, on this basis, only the QTL on chromosome 9, the nonsignificant QTL on chromosome 13 and the second QTL on chromosome 5 received no corroborating evidence. Several of the putative QTL detected here confirm QTL identified in other studies. Zhang et al. (1998) completed a genome-wide scan and detected a QTL affecting milk yield and protein percent on chromosome 3 of Holsteins. Combining the information from Zhang et al. (1998) with a previous publication of marker information for the same experiment (Georges et al., 1995), and comparing this to the MARC and other published maps available at http://www.ri.bbsrc.ac.uk/cgi-bin/ mapviewer?species=cattle, indicates that the estimated position of the QTL detected on chromosome 3 in Zhang et al. (1998) would lie somewhere between 30 and 50 cM on the map used here (Table 3 and Figure 1). It thus seems quite likely that this is the same QTL as that affecting milk yield, fat, and protein percentage detected here. In another genome-wide scan of Holstein cattle Heyen et al. (1999), up to three QTL variously affecting milk fat and protein yield and fat and protein percentage were detected in three sire families in the region between 0 and about 50 cM on our map for chromosome 3. The same study also located a QTL affecting fat percentage on chromosome 5 located at about 90 cM on Journal of Dairy Science Vol. 84, No. 6, 2001

our map. This could be the same as the QTL affecting fat yield located at 63 cM in sire family 3 in our study, but no estimate of effects on other traits were provided in Heyen et al. (1999) making it more difficult to judge the likelihood of this being the same QTL. The QTL affecting milk, fat, and protein yield detected by Zhang et al. (1998) on chromosome 9 would map to about 40 to 50 cM on the map used here. This is consistent with the QTL with effects similar to those we detected in family 2 located at about 40 to 60 cM. A study of chromosome 9 of Finnish Ayrshires (Vilkki et al., 1997) also found suggestive evidence for the presence of a QTL affecting milk and protein yield located at about 35 cM on our chromosome 9 map. On chromosome 17, the QTL affecting milk yield detected by Zhang et al. (1998) would map at the end of our map, around 37 cM, which is consistent with the QTL with similar effects that we located here. In a separate study of American Holsteins (Ashwell et al., 1998), an association between marker BM8125 (located at 12 cM on chromosome 17 on our map) and fat percent was detected. This is consistent with the fat percent QTL we detected in family 2 at an estimated location of 0 cM. On chromosome 20, the QTL affecting fat and protein percent detected by Zhang et al. (1998) and confirmed by Arranz et al. (1998) would map to about 40 cM on our map. While the location is consistent with both of the putative QTL we detected on chromosome 20, the effects are consistent only with the QTL detected in family 3, located at about 64 cM on our map. We were unable to determine the exact location of the QTL affecting fat yield detected by Zhang et al. (1998) on chromosome 23. One of the markers flanking the QTL location in Zhang et al. (1998) was MGTG7 which would lie at about 15 cM on our map, but it is not clear whether their estimated QTL position lies proximally or distally from MGTG7 on our map. The effect of the QTL is certainly consistent with the QTL we detected in the general region of 5 to 45 cM in three sire families. In a study of chromosome 23 of Finnish Ayrshires (Elo et al., 1999), a possible QTL affecting protein percent was located, but there was no suggestion that the QTL detected in the present study had any effect on protein percent. Finally, the QTL affecting fat percent detected by Zhang et al. (1998) on chromosome 26 would map to somewhere between 20 and 25 cM on our map, which is consistent with the estimated location of 5 to 35 cM for the QTL affecting milk and fat yield and fat percent in two sire families here. Overall, there is consistency between the results of Zhang et al. (1998) and those presented here. The two studies had 10 chromosomes in common. On chromosome 5, neither study found a significant effect, while on the remaining nine chromosomes, seven provided

QTL FOR PRODUCTION TRAITS OF HOLSTEIN CATTLE

estimates of QTL that were consistent in location and general effect. A specific study of chromosome 9 in Finnish Ayrshires located a similar effect as here and in Zhang et al. (1998). In contrast, the genome-wide study of Ashwell and Van Tassell (1999) shared 10 chromosomes with the present study and 29 with Zhang et al. (1998), but presented no estimates of QTL in common with either study. Similarly, a genome-wide scan of US and Israeli Holsteins (Heyen et al., 1999) also shared only one QTL clearly in common with our study and had only a small number in common with Zhang et al. (1998). Interestingly, an earlier analysis (Ashwell et al., 1998) of a subset of the data included in Ashwell and Van Tassell (1999) found effects of chromosome 17 consistent with both here and in Zhang et al. (1998) and an effect on chromosome 14 consistent with an effect noted by Zhang et al. (1998) and Heyen et al. (1999). Since Ashwell and Van Tassell (1999) used much more stringent criteria for significance than in their previous study (Ashwell et al., 1998), this points to the dangers of missing real QTL when applying too high a stringency for significance. In general, no matter how stringent the chosen significance level, rarely will an initial test for the presence of a QTL be sufficient to provide full confidence that a QTL effect is real. Comparison of effects across studies is a powerful tool for confirmation of QTL, but is hampered by the current focus on only publishing the most highly significant results. Authors should be encouraged both to present a broader range of results and to deposit more information into web-based databases to allow broader comparisons across studies. In conclusion, notwithstanding the large errors attached to putative QTL locations and affects detected here, the various lines of evidence suggest that the majority of these are real QTL, which could be useful in future marker based selection schemes. ACKNOWLEDGMENTS This research was funded by the Cattle Breeding Research Council (Canada), the National Research Council Industrial Research Assistance Program (Canada), the Natural Sciences and Engineering Research Council (Canada), and the Ontario Ministry of Agriculture, Food and Rural Affairs. Daughter yield deviations and pedigree data were kindly supplied by Janusz Jamrozik of the Canadian Dairy Network. S. Knott, P. Visscher, C. Haley, D. Nielson, and P. Green kindly supplied us with copies of software. We thank L. R. Schaeffer for statistical advice and J. L. Atchison, X. Wu, and M. Lu for laboratory work.

1523

REFERENCES Arranz, J. J., W. Coppieters, P. Berzi, N. Cambisano, B. Grisart, L. Karim, F. Marcq, L. Moreau, C. Mezer, J. Riquet, P. Simon, P. Vanmanshoven, D. Wagenaar, and M. Georges. 1998. A QTL affecting milk yield and composition maps to bovine chromosome 20: a confirmation. Anim. Genet. 29:107–151. Ashwell, M. S., Y. Da, C. P. Van Tassell, P. M. VanRaden, R. H. Miller, and C. E. Rexroad, Jr. 1998. Detection of putative loci affecting milk production and composition, health, and type traits in a United States Holstein population. J. Dairy Sci. 81:3309– 3314. Ashwell, M. S., and C. P. Van Tassell. 1999. Detection of putative loci affecting milk, health and type traits in a US holstein population using 70 microsatellite markers in a genome scan. J. Dairy Sci. 82:2497–2502. Barendse, W., D. Vaiman, S. J. Kemp, Y. Sugimoto, S. M. Armitage, J. L. Williams, H. S. Sun, A. Eggen, M. Agaba, S. A. Aleyasin, M. Band, M. D. Bishop, J. Buitkamp, K. Byrne, F. Collins, L. Cooper, W. Coppettiers, B. Denys, R. D. Drinkwater, K. Easterday, C. Elduque, S. Ennis, G. Erhardt, L. Ferretti, N. Flavin, Q. Gao, M. Georges, R. Gurung, B. Harlizius, G. Hawkins, J. Hetzel, T. Hirano, D. Hulme, C. Jorgensen, M. Kessler, B. W. Kirkpatrick, B. Konfortov, S. Kostia, C. Kuhn, J. A. Lenstra, H. Leveziel, H. A. Lewin, B. Leyhe, L. Lil, I. Martin Burriel, R. A. McGraw, J. R. Miller, D. E. Moody, S. S. Moore, S. Nakane, I. J. Nijman, I. Olsaker, D. Pomp, A. Rando, M. Ron, A. Shalom, A. J. Teale, U. Thieven, B.G.D. Urquhart, D.-I. Vage, A. Van de Weghe, S. Varvio, R. Velmala, J. Vilkki, R. Weikard, C. Woodside, J. E. Womack, M. Zanotti, and P. Zaragoza. 1997. A medium-density genetic linkage map of the bovine genome. Mamm. Genome 8:21–28. Bishop, M. D., S. M. Kappes, J. W. Keele, R. T. Stone, S. L. Sunden, G. A. Hawkins, S. S. Toldo, R. Fries, M. D. Grosz, and J. Yoo. 1994. A genetic linkage map for cattle. Genetics 136:619–639. Doerge, R. W., and G. A. Churchill. 1996. Permutation tests for multiple loci affecting a quantitative character. Genetics 142:285–294. Elo, K. T., J. Vilkki, D. J. de Koning, R. J. Velmala, and A. V. MakiTanila. 1999. A quantitative trait locus for live weight maps to bovine chromosome 23. Mamm. Genome 10:831–835. Georges, M., D. Nielsen, M. Mackinnon, A. Mishra, R. Okimoto, A. T. Pasquino, L. S. Sargeant, A. Sorensen, M. R. Steele, and X. Zhao. 1995. Mapping quantitative trait loci controlling milk production in dairy cattle by exploiting progeny testing. Genetics 139:907–920. Gomez-Raya, L., and J. P. Gibson. 1993. Within-family selection at an otherwise unselected locus in dairy cattle. Genome 36:433–439. Heyen, D. W., J. I. Weller, M. Ron, M. Band, J. E. Beever, E. Feldmesser, Y. Da, G. R. Wiggans, P. M. VanRaden, and H. A. Lewin. 1999. A genome scan for QTL influencing milk production and health traits in dairy cattle. Physiological Genomics 1:165–175. Kashi, Y., E. Hallerman, and M. Soller. 1990. Marker-assisted selection of candidate bulls for progeny testing programmes. Anim. Prod. 51:63–74. Knott, S. A., J. M. Elsen, and C. S. Haley. 1996. Methods for multiplemarker mapping of quantitative trait loci in half-sib populations. Theor. Appl. Genet. 93:71–80. Lipkin, E., M. O. Mosig, A. Darvasi, E. Ezra, A. Shalom, A. Friedmann, and M. Soller. 1998. Quantitative trait locus mapping in dairy cattle by means of selective milk DNA pooling using dinucleotide microsatellite markers: analysis of milk protein percentage. Genetics 149:1557–1567. Litt, M., and J. A. Luty. 1989. A hypervariable microsatellite revealed by in vitro amplification of a dinucleotide repeat within the cardiac muscle actin gene. Am. J. Hum. Genet. 44:397–401. Nadesalingam, J., Y. Plante, and J. P. Gibson. 2001. Detection of quantitative trait loci on chromosomes 1 and 6 affecting milk production traits in Canadian Holsteins. Mamm. Genome 12:27–31. Nadesalingham, J. 1999. Mapping and detection of quantitative trait loci (QTL) for milk production traits based on linkage analysis using microsatellite markers in Canadian Holstein cattle. PhD. Guelph, Guelph. Journal of Dairy Science Vol. 84, No. 6, 2001

1524

PLANTE ET AL.

Spelman, R. J., W. Coppieters, L. Karim, J. A. van Arendonk, and H. Bovenhuis. 1996. Quantitative trait loci analysis for five milk production traits on chromosome six in the Dutch HolsteinFriesian population. Genetics 144:1799–1808. VanRaden, P. M., and G. R. Wiggans. 1991. Derivation, calculation and use of National Animal Model Information. J. Dairy Sci. 74:2737–2746. Velmala, R. J., H. J. Vilkki, K. T. Elo, D. J. de Koning, and A. V. Maki-Tanila. 1999. A search for quantitative trait loci for milk production traits on chromosome 6 in Finnish Ayrshire cattle. Anim. Genet. 30:136–143.

Journal of Dairy Science Vol. 84, No. 6, 2001

Vilkki, H. J., D. J. de Koning, K. Elo, R. Velmala, and A. MakiTanila. 1997. Multiple marker mapping of quantitative trait loci of Finnish dairy cattle by regression. J. Dairy Sci. 80:198–204. Weller, J. I., Y. Kashi, and M. Soller. 1990. Power of daughter and granddaughter designs for determining linkage between marker loci and quantitative trait loci in dairy cattle. J. Dairy Sci. 73:2525–2537. Zhang, Q., D. Boichard, I. Hoeschele, C. Ernst, A. Eggen, B. Murkve, M. Pfister-Genskow, L. A. Witte, F. E. Grignola, P. Uimari, G. Thaller, and M. D. Bishop. 1998. Mapping quantitative trait loci for milk production and health of dairy cattle in a large outbred pedigree. Genetics 149:1959–1973.