Copyright is owned by the Author of the thesis.  Permission is given for 
a copy to be downloaded by an individual for the purpose of research and 
private study only.  The thesis may not be reproduced elsewhere without 
the permission of the Author. 
 
 
Genetics of flystrike, dagginess and 
associated traits in New Zealand dual-purpose 
sheep  
 
 
 
 
A thesis presented in partial fulfilment of the requirements 
for the degree of  
 
 
 
 
Doctor of Philosophy  
in  
Animal Science  
 
 
 
at Massey University, Palmerston North, New Zealand 
 
 
 
Natalie Kathleen Pickering 
 
2013 
 
 
Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
ii 
 
 
 
 
 
 
 
 
?The greatest benefit of dag control or elimination may come from reduced fly 
strike? V. J. Mackereth (1983), an astute breeder. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Abstract 
iii 
 
Abstract 
Pickering, N. K. (2013). Genetics of flystrike, dagginess and associated traits in 
New Zealand dual-purpose sheep. PhD thesis, Massey University, Palmerston North, 
New Zealand. 
  
A literature review identified breech bareness, dagginess and fibre traits as potential 
indirect indicator traits for flystrike. Dagginess (faecal accumulation) had the greatest 
potential as an indirect indicator, and has been identified as an important trait itself. 
Therefore flystrike and dagginess were investigated for their associations with fibre and 
production traits. A genome-wide association study (GWAS) was performed to identify 
regions under selection and associated with these traits. Finally, a genomic selection 
(GS) analysis was performed for dagginess and dual-purpose production traits to 
estimate molecular breeding values (MBVs) and to determine their impact on the New 
Zealand dual-purpose selection index.  
Heritability, genetic and phenotypic parameter estimations were performed on a 
flystrike case-control dataset collected over 2 years. Flystrike had a heritability of 0.37, 
and high genetic and phenotypic correlations with dag score and a high genetic 
correlation with the coefficient of variation of fibre diameter. A similar analysis was 
performed on an existing New Zealand sheep industry dataset of about 2 million 
pedigree-recorded animals born between 1990 and 2008. The heritability for dag score 
at 3 and 8 months (DAG3, DAG8) was 0.34 and 0.31 respectively. There were low or 
nil genetic and phenotypic correlations of DAG3 and DAG8 with the other standard live 
weight, fleece weight, reproduction and faecal egg count production traits or breech 
bareness, fibre and wool traits.  
A GWAS performed on an industry dataset of 8,705 genotyped animals, using 
phenotype information on about 3 million pedigree-recorded animals, identified regions 
on chromosome 6 and 15 associated with DAG3 and DAG8. The lambs from the 
flystrike case-control dataset with SNPs imputed from 5K to 50K identified a number of 
immune, diarrhoea and wool/hair growth genes associated with flystrike, dag score and 
fibre traits in a GWAS. There were no similarities in the genes identified in the industry 
or case-control GWAS; however, the SNP on chromosome 15 was re-identified in the 
GS analysis for DAG8. The GS analysis showed that genomic predictions can be 
Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
iv 
 
performed for DAG3 and DAG8 and that using MBVs and modifying generation 
interval can increase the rate of the genetic gain of the dual-purpose index by 84% per 
year. 
 
 
 
 
 
 
 
 
 
 
 
Acknowledgments 
v 
 
Acknowledgements 
At least 4 years ago and maybe back the full 28 years, I have had the guidance of a 
forward thinking father who did not want to crutch another sheep. Also along the last 6 
years I have had the opportunity to work with another forward thinking and supportive, 
first boss then supervisor, John McEwan. Together I have them to thank for one 
suggesting the topic and the other in moulding my scientific skills so as to answer the 
question ?Can I breed for no dags and flystrike?? 
There are a number of people who I?ve had the opportunity to work with and learn 
off during this experience. Firstly my full contingent of supervisors: Hugh Blair, 
Rebecca Hickson from Massey University and Tricia Johnson and John McEwan from 
AgResearch, Invermay. Thank you for your contributions to shaping me, my 
experiments, and my thesis. Also for teaching me ?past and present? tense. I would also 
like to thank Ken Dodds and Benoit Auvray for their guidance and invaluable help with 
the statistical component of the project, without them the languages, programs and 
methods such as R, SAS, ASReml, genomic selection and GWAS would still be foreign 
outer space ?things?! I?ve also had the pleasure of sharing my workspace with the 
Invermay Animal Genomics team, some have come and gone since I first started at 
AgResearch, but all have contributed to the great atmosphere which has been a pleasure 
to work in.   
Secondly, the financial support of Ovita and AgResearch for funding the research 
and for the provision of my doctoral stipend. Also the IVABS Travel Award, Early 
Career Scientist Travel Bursary, and the Charles Elgar Scholarship awards which have 
allowed me to attend and/or present my work at 9th World Congress on Genetics 
Applied to Livestock Production (2010), The Applied Genomics for Sustainable 
Livestock Breeding Conference (2011), New Zealand Society of Animal Production 
(2012), and the 32nd and 33rd International Society of Animal Genetics (2010, 2012) 
conferences. 
Thirdly, this project would not be possible without the support of the farmers that 
have allowed me access to their flocks via SIL and have allowed me to look at ~ 30,000 
dirty and clean lamb bottoms. Also the efforts of those that collected samples and 
measurements of flystruck lambs for me, a not very pleasant job, were very much 
Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
vi 
 
appreciated. A big thank you for those that provide me a bed and great home cooking 
during my road trips around the country collecting measurements; these provided some 
of my most memorable moments during my project.  
A big thank you to my family; Mum, Dad and brothers for their amazing support 
throughout my life, and helping me through the PhD by being a willing participant in 
the case-control study. Finally to Jason, who has been a rock for me for the past 3 years, 
putting up with my stressed out moments and encouraging me to try new things like 
scuba diving.  
I thank you all. 
 
 
 
 
 
 
 
Table of contents 
vii 
 
Table of contents 
Abstract           iii 
Acknowledgements           v 
Table of contents         vii 
List of tables          viii 
List of figures         xiii 
List of abbreviations         xix 
Introduction          xxi 
Chapter 1: Review of literature        25 
Chapter 2: Case-control experiment: estimation of genetic parameters and a summary 
of two flystrike seasons         71 
Chapter 3: Genetic parameters for production traits in New Zealand dual-purpose 
sheep, with an emphasis on dagginess       94 
Chapter 4: Genetic relationships between dagginess, breech bareness and wool traits in 
New Zealand dual-purpose sheep       113 
Chapter 5: Evaluation of sampling, paternity parentage and imputation  133 
Chapter 6: Genome-wide association study: flystrike case ? controls  160 
Chapter 7: Genome-wide association study: SIL industry data including bare breech 
and fibre traits         187 
Chapter 8: The impact of genomic selection on dual-purpose selection index including 
dagginess          228 
Chapter 9: Concluding discussion       271 
References          285 
Appendices          318 
 
 Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
 
viii 
 
List of tables 
Table 1.1: Distribution of strikes, in 3 separate studies from the United Kingdom (UK), 
New Zealand (NZ), and Australia (AUS)       28 
Table 1.2: Heritability (h2) of potential indirect selection traits, and their genetic, 
phenotypic and between flock correlations with flystrike in Australian Merino  57 
Table 2.1: Number of flystrike cases (n), incidence rate (I, %), proportion of flystrike 
located on the breech (B, %) and preventative insecticide use for lambs born 2009 
and 2010 per flock         83 
Table 2.2: Summary of strikes by area for each flystrike season    84 
Table 2.3: Least square means of each trait for cases and controls    84 
Table 2.4: Final mixed models and fixed effects used for individual trait analysis  85 
Table 2.5: ASReml estimates of heritabilities (diagonal), phenotypic (above diagonal), 
genotypic (below diagonal) correlations, and phenotypic standard deviation (?p) ? 
s.e            87 
Table 2.6: Insecticide use by farmers (A to K) for the prevention and treatment of 
flystrike. IGR: insect growth regulator, OP: organophosphates, SP: synthetic 
pyrethoids           88 
Table 3.1: Final mixed models and fixed effects used for individual trait analysis  99 
Table 3.2: Least square means and ANOVA summary for traits   100 
Table 3.3: Estimates of heritabilities (diagonal), phenotypic (above diagonal), 
genotypic (below diagonal) correlations, and phenotypic standard deviations (?p) 
and repeatability estimates (last row) ? s.e     102 
Table 3.4: Estimates of permanent environmental correlations (above diagonal) and 
genetic plus permanent environmental correlations (below diagonal) ? s.e 103 
Table 3.5: Source estimates of current SIL parameters    106 
Table 3.6: Summary of estimated breeding values (EBVs) results for 3 flocks run with 
the current Sheep Improvement Limited (SIL) and new dag models  109 
Table 4.1: Summary of 2009/2010 progeny dataset before data cleaning, transformation 
and addition of dam information.       118 
Table 4.2: Final mixed models and fixed effects including contemporary groups used 
for individual trait analysis       122 
List of tables 
ix 
 
Table 4.3: Means and ANOVA summary for traits after cleaning, transformation and 
addition of dam information for WWT and LWAU    123 
Table 4.4: Estimates of heritabilities (diagonal), phenotypic (above diagonal), 
genotypic (below diagonal) correlations, and phenotypic standard deviations (?p) ? 
s.e               126-127 
Table 5.1: Sample types collected as part of the case-control experiment  136 
Table 5.2: Average, standard deviation (SD), minimum (Min) and maximum (Max) 
amount (?g) of DNA from tissue samples successfully extracted  142 
Table 5.3: Sample type and number of animals that were genotyped by the Sequenom? 
assay          143 
Table 5.4: Genotyping success rate of the flystrike case-controls run on the 5K SNP 
Chip and the sires run on the 50K SNP Chip     146 
Table 5.5: Average proportion of errors for each imputed SNP and imputed progeny for 
each program option.        148 
Table 5.6: Comparison of paternity testing programs. Number of paternity assigned 
progeny per program on the diagonal (percentage). Below the diagonal actual values 
of paternity tested progeny with the same result in 2 tests (percentage), and number 
of total assigned progeny per SNP platform, number of unassigned progeny per test 
(percentage) and number of progeny that passed quality control on the 2 platforms; 
Sequenom (Cervus and Partial Pedigree) and 5K SNP Chip   149 
Table 5.7: Computational time taken to impute each chromosome (Chr) in days hours 
(hrs) and minutes (mins)        151 
Table 6.1: Power of analysis for given sample size (n), and parameters for a dominant 
model          165 
Table 6.2: Power of analysis for given sample size (n), and parameters for a recessive 
model          166 
Table 6.3: Estimate of lambda (slope), and their standard error (s.e.) of the linear 
regression of the observed -log10(P) on the expected -log10(P) (QQplot) for each trait
           168 
Table 6.4: The trait, number of animals (n), marker name, chromosome (Chr) and 
chromosome position (base pairs, bp) of the best 5 SNPs for the traits flystrike (as 
log10(scaled flystrike +1)), dag score (DAG), breech bareness (BBREECH), wool 
bulk (BULK) and wool length (LENGTH). The P value and -log10(P) are shown for 
each marker         175 
Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
x 
 
Table 6.5: The trait, number of animals (n), marker name, chromosome (Chr) and 
chromosome position (base pairs, bp) of the best 5 SNPs for the fibre traits; mean 
fibre diameter (MFD), standard deviation of MFD (FDSD), coefficient of variation 
of MFD (FDCV), curvature (CURV) and proportion medullation (MED%). The P 
value and -log10(P) are shown for each marker     176 
Table 6.6: List of SNPs identified with immune, inflammatory response and diarrhoea 
roles (greyed) for the traits flystrike and dag score, which reached nominal 
significant threshold P<0.001       178 
Table 6.7: List of SNPs identified with wool/hair roles (greyed) for breech bareness 
(BBREECH), mean fibre diameter standard deviation (FDSD) and proportion 
medullation (MED%) for chromosome 3, 7 and 8 which reached the nominal 
significant threshold P<0.001       179 
Table 6.8: List of SNPs identified with wool/hair roles (greyed) for mean fibre diameter 
standard deviation (FDSD) and proportion medullation (MED%) on chromosome 13 
and 23 which reached the nominal significant threshold P<0.001  180 
Table 7.1: Summary table of the number of animals (n), heritability (h2), reliability cut 
off (0.8h2), average and maximum reliability (Av Rel, Max Rel) of breeding values 
used in the analysis        196 
Table 7.2: The breed genotypic variance explained by the first 6 principle components 
(PC) for each trait         197 
Table 7.3: Estimate of lambda (slope), and their standard error (s.e.) of the linear 
regression of the observed -log10(P) on the expected -log10(P) (QQplot) for each trait
           199 
Table 7.4: The trait, number of animals (n), marker name, chromosome (Chr) and 
chromosome position (base pairs, bp) of the best 3 SNPs for traits with SIL 
calculated BVs. The P value and -log10(P) are shown for each marker  200 
Table 7.5: The trait, number of animals (n), marker name, chromosome (Chr) and 
chromosome position (base pairs, bp) of the top 3 SNPs for the wool and fibre traits. 
The P value and -log10(P) are shown for each marker    204 
Table 7.6: Candidate genes within 100kbp of the best SNPs for the SIL traits, by 
chromosome (Chr)        205 
Table 7.7: Correlation (r) of -log10(P) values between case-control GWAS (C6, Chapter 
6) and industry GWAS (C7, this Chapter)     207 
List of tables 
xi 
 
Table 7.8: List of significant peaks and the candidate genes within the regions. The 
greyed genes are discussed in more detail           210-211 
Table 8.1: The year of birth of the first animals placed in the validation set and number 
(n) of animals in training and validation sets for each breed   234 
Table 8.2: Phenotypic standard deviation (?p), heritability (h
2), repeatability (Rep), 
economic weights (EW), and number of records used for each trait in the dual-
purpose selection index for selecting a 2 year old ram    239 
Table 8.3: The breed genotypic variance explained by the first 6 principal components 
(PC) for each trait         241 
Table 8.4: Variance components; additive (?u
2) and corrected residual (?e
2) variance, 
and the heritabilities used; (h2) from Sheep Improvement Limited (SIL) and the 
effective h2 for each trait        243 
Table 8.5:  The weighted correlations from G1 and G6z analysis, between mBVs and 
dependent variables in the 7 validation breeds     244 
Table 8.6: Accuracies of the weighted correlation between MBV and dependent 
variable (rA), the weighted average individual accuracy (rI) and the combined 
accuracy (rC) calculated for the 16 traits in the 7 validation breeds  245 
Table 8.7: The lower limit of the 90% confidence interval for the weighted correlation 
between MBV and dependent variable (rA),  for all traits and all 7 validation breeds
           246 
Table 8.8: Proportion of variance explained by MBVs in 7 validation breeds 247 
Table 8.9: The k values for traits in the 7 validation breeds     248 
Table 8.10: The best 5 SNPs for the traits: weaning weight (WWT), fleece weight at 12 
months (FW12), dag score at 3 and 8 months (DAG3, DAG8), with chromosome 
(Chr), Chr position (base pairs, bp), and the P value and -log10(P) value 249 
Table 8.11: The correlation (r), number of SNPs in top 20 for both analyses (Top20) 
and summary of the -log10(P) values from the 2 analyses (GWAS and GS) 250 
Table 8.12: Estimate of lambda (slope), and their standard error (s.e.) of the linear 
regression of the observed -log10(P) on the expected -log10(P) (QQ plot) for each 
trait          254 
Table 8.13: The heritability (h2), repeatability (rep) and genomic selection accuracy 
(GS acc), effective number of progeny (n prog) and their single trait selection 
accuracy (Seln acc) estimated from the breed combined-accuracy (rC) and the 
individual accuracy (rI)        256 
Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
xii 
 
Table 8.14:  Response per selection round per trait (change (?) in units and dollars ($)) 
and accuracies, and overall index (standard deviation (SD) index in dollars) and 
accuracy for a dual-purpose Romney index     258 
Table 8.15: Response per selection round (SD index) and accuracy (acc), conversion 
factor (i/L, selection intensity / generation interval), and rate of genetic gain (?G), 
for each scenario         259 
Table 8.16: Number of animals to genotyped to reach accuracy of ~0.70 if have own 
measurement or 126 progeny measurements     262 
 
List of figures 
xiii 
 
List of figures 
Figure 1.1: Images of the 4 main fly species with larvae inset (not to scale): L. cuprina 
(A), L. sericata (B), C. stygia (C) and C. rufifacies (D). The two Lucilia species are 
separated by the green colouration of the fore femora of L. cuprina and the 
increased number of hairs of the humeral calli and notopleuron area of L. sericata 
(Lang et al., 2001)          28 
Figure 1.2: Classification of sheep body regions in strike identification. 1: Breech, 2: 
Body, 3: Shoulders, 4: Head, 5: Belly (and Pizzle in males), 6: Foot. Modified from 
French et al.(1995)         29 
Figure 1.3: Lifecycle of the fly        30 
Figure 1.4: Dag scoring system: 0 no dags to 5 most daggy. As proposed by Sheep 
Improvement Limited (www.sil.co.nz/Files/Tech-Notes/DagScoreV2.aspx)  46 
Figure 1.5: Breech bareness scoring system: 1: fleece cover extends to margins of anus, 
to 5: an extensive bare area either side of the anus. As proposed by Sheep 
Improvement Limited (www.sil.co.nz/getdoc/a9bb121b-2016-4b34-a163-399ae 
4b28471/Doc-ID-000010-GW-Bare-Points-sheep.aspx)     52 
Figure 2.1: The 6-month flystrike season overlapped by the timing of weaning and live 
weight at 6 months (LW6) recording, when the majority of flystruck lambs were 
observed and measured by the farmer       74 
Figure 2.2: Map of experimental sites. Properties involved in case-control test; 
commercial farmers ( ) and Breeders ( )      75 
Figure 2.3: Classification of sheep body regions in strike identification. 1: Breech, 2: 
Body, 3: Shoulders, 4: Head, 5: Belly (and Pizzle in males), 6: Foot. Modified from 
French et al. (1995)         76 
Figure 2.4: Total rainfall (mm) from July 2009 to 30th June 2010 (top left), and over the 
flystrike season, October 2009 to 31st March 2010 (top right), July 2010 to 30th June 
2011 (bottom left), and over the flystrike season, October 2010 to 31st March 2011 
(bottom right)          81 
Figure 2.5: Average maximum (red line) and minimum (blue line) temperature (?C) for 
the 2009/10 season (left) and 2010/11 season (right) across all farms    82 
Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
xiv 
 
Figure 3.1: Plot of standard deviation versus mean for contemporary groups (flock-
year-sex) for new dag score at 3 months model (A) and current Sheep Improvement 
Limited dagginess model (B), after scaling for all 3 flocks tested  110 
Figure 3.2: Plot of current Sheep Improvement Limited dagginess EBV value (x axis) 
versus new dag score at 3 months (DAG3; A), and at 8 months (DAG8; B) EBV 
values (y axis) for flock A       111 
Figure 4.1: Map of experimental sites; properties involved in the progeny test (?)
           116 
Figure 5.1: Frequency distribution of DNA concentrations (?g) from tissue extractions. 
Arrows indicate 1 spleen sample at 322 ?g and 389 ?g.    142 
Figure 5.2: Gel picture of DNA from TypiFix? ear tissue extractions. Samples were 
run in duplicate, on a 1.5% agarose gel with 1kb plus ladder   143 
Figure 5.3: Frequency distribution of call rate (%) of all samples genotyped through the 
Sequenom? assay derived from the 115 SNPs genotyped   144 
Figure 5.4: Frequency distribution of call rate (%) of all 115 Sequenom? SNPs across 
all animals          144 
Figure 5.5: Minor allele frequency (MAF) of SNPs genotyped over the progeny on the 
5K SNP Chip. The average MAF noted by arrow    146 
Figure 5.6:  Minor allele frequency (MAF) of SNPs genotyped over the sires on the 
50K SNP Chip. The average MAF noted by arrow    147 
Figure 5.7: Density graph of error rates for Option1: animals (A), SNPs (B), and 
Option 5: animals (C), SNPs (D). Option 1: LinkPHASE, DAGPHASE and 
BEAGLEv2.1.3: 287 progeny and 129 sires including pedigree information. Option 
5: BEAGLEv3.0.4: 227 progeny and 129 sires (?pairs? file) plus the 60 unrelated 
progeny and 1130 unrelated animals (unrelated file)    148 
Figure 5.8: Comparison of error rate for the imputed progeny per SNP with the minor 
allele frequency in the imputed data set for each SNP (0-1 line, red). Option 1: 
LinkPHASE, DAGPHASE and BEAGLEv2.1.3: 287 progeny and 129 sires 
including pedigree information (left). Option 5: BEAGLEv3.0.4: 227 progeny and 
129 sires (?pairs? file) plus the 60 unrelated progeny and 1130 unrelated animals 
(unrelated file) (right)        149 
Figure 5.9: Minor allele frequencies (MAF) for unrelated best guess (A), unrelated 
genotype probabilities (B), pairs (C), and the correlation between best guess and 
genotype probabilities MAF       152 
List of figures 
xv 
 
Figure 5.10: Diagonal values of the G matrix using the genotype probabilities for the 
253 unrelated progeny (left of red line), and the pairs phased genotypes for the 
paternity assigned progeny (right of red line)     153 
Figure 5.11: Diagonal values of the G matrix using the best guess genotypes for the 
253 unrelated progeny (left of red line), and the pairs phased genotypes for the 
paternity assigned progeny (right of red line)     154 
Figure 5.12: The comparison of the diagonal values of the G matrix from the genotype 
probabilities and the best guess genotypes. Including the paternity pairs phased 
genotypes          154 
Figure 6.1: Plot of the power for a dominant (blue) and recessive (red) model against 
the heritability (h2) of the QTL       166 
Figure 6.2: Plot of the first 4 principal components (PC) for flystrike case and control 
derived from genotype information. Coloured by farmer, see Chapter 2. Farmer A 
(orange), B (yellow), C (Dark blue), D (green), E (blue), F (pink), G (red), H (light 
grey), I (grey) and K (brown)       167 
Figure 6.3: QQ plots for flystrike (left) and dag score (right) -log10(P) values. The 0-1 
line is in black and the linear regression in red     169 
Figure 6.4: Graph of -log10(P) values of SNPs for flystrike, ordered on ovine genome 
v1 map, P<0.001 (red dash line)       170 
Figure 6.5: Graph of -log10(P) values of SNPs for dag score, ordered on ovine genome 
v1 map, P<0.0001 (solid red line), P<0.001 (red dash line)   170 
Figure 6.6: Graph of -log10(P) values of SNPs for breech bareness score, ordered on 
Ovine genome v1 map, P<0.0001 (solid red line), P<0.001 (red dash line) 171 
Figure 6.7: Graph of -log10(P) values of SNPs for wool bulk, ordered on ovine genome 
v1 map, P<0.001 (red dash line)       171 
Figure 6.8: Graph of -log10(P) values of SNPs for wool length, ordered on ovine 
genome v1 map, P<0.0001 (solid red line), P<0.001 (red dash line)  172 
Figure 6.9: Graph of -log10(P) values of SNPs for mean fibre deviation, ordered on 
ovine genome v1 map, P<0.001 (red dash line)     172 
Figure 6.10: Graph of -log10(P) values of SNPs for standard deviation of mean fibre 
deviation, ordered on ovine genome v1 map, P<0.001 (red dash line)  173 
Figure 6.11: Graph of -log10(P) values of SNPs for coefficient of variation of mean 
fibre diameter score, ordered on ovine genome v1 map, P<0.0001 (solid red line), 
P<0.001 (red dash line)        173 
Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
xvi 
 
Figure 6.12: Graph of -log10(P) values of SNPs for curvature, ordered on ovine genome 
v1 map, P<0.0001 (solid red line), P<0.001 (red dash line)   174 
Figure 6.13: Graph of -log10(P) values of SNPs for proportion medullation score, 
ordered on ovine genome v1 map, P<0.00001 (black dash line), P<0.0001 (solid red 
line), P<0.001 (red dash line)       174 
Figure 6.14: Plot of the power for a dominant (blue/green) and recessive (red/purple) 
model when 1,000 (diamond) or 667 (triangle) cases are collected against the 
heritability (h2) of the QTL       182 
Figure 7.1: Minor allele frequency (MAF) for the Ovine 50K SNP chip, following 
removal, for quality control checks of the 8579 animals retained for GWAS. Arrow 
indicates mean MAF        196 
Figure 7.2: The first 4 principal components (PC) calculated from the G matrix for 
weaning weight. Romney (blue), Coopworth (green), Texel (yellow), Perendale 
(purple) and others (grey)        198 
Figure 7.3: QQ plot for weaning weight (A), fleece weight at 12 months (B), dag score 
at 3 months (C) and 8 months (D) -log10(P) values. The 0-1 line is in black and the 
linear regression in red        199 
Figure 7.4: Manhattan plot of -log10(P) values of SNPs for weaning weight. Ordered on 
the ovine genome v2 map, P<0.001 (red dash line)    201 
Figure 7.5: Manhattan plot of -log10(P) values of SNPs for fleece weight at 12 months. 
Ordered on the ovine genome v2 map      201 
Figure 7.6: Manhattan plot of -log10(P) values of SNPs for dag score at 3 months. 
Ordered on the ovine genome v2 map, P<0.0001 (solid red line), P<0.001 (red dash 
line)          202 
Figure 7.7: Manhattan plot of -log10(P) values of SNPs for dag score at 8 months. 
Ordered on the ovine genome v2.0 map, P<0.0001 (solid red line), P<0.001 (red 
dash line)          202 
Figure 7.8: Plot of -log10(P) values for individual SNPs from the case-control GWAS 
(y axis) for dag score versus the industry GWAS (x axis) for dag score at 3 months 
(DAG3, left) and at 8 months (DAG8, right). The 0-1 line is plotted in red 207 
Figure 7.9: Animals clustered on the basis of principal components 1 and 2 (PC 1, PC 
2). Romney and Arapawa breeds shown in blue and green respectively  208 
List of figures 
xvii 
 
Figure 7.10: A Manhattan plot of the moving window of 5 (WIN5) FST values between 
the Arapawa and high dag score breeding value Romneys. Ordered on the ovine 
genome v2 map, WIN5 FST = 0.3 (red dash line)     209 
Figure 7.11: Schematic of significant region on chromosome 2, SNPs (diamonds), 
scaffold (green lines), and bovine ref sequences (black line)   211 
Figure 7.12: Schematic of significant region on chromosome 10, SNPs (diamonds), 
scaffold (green lines) and bovine ref sequences (black line)   211 
Figure 7.13: Schematic of significant region 1 on chromosome 13, SNPs (diamonds), 
scaffold (green lines) and bovine ref sequences (black line)   212 
Figure 7.14: Schematic of significant region 2 on chromosome 13, SNPs (diamonds), 
scaffold (green lines) and bovine ref sequences (black line)   212 
Figure 7.15: Schematic of significant region 3 on chromosome 13, SNPs (diamonds), 
scaffold (green lines) and bovine ref sequences (black line)   213 
Figure 7.16: Schematic of significant region 4 on chromosome 13, SNPs (diamonds), 
scaffold (green lines) and bovine ref sequences (black line)   213 
Figure 7.17: Schematic of significant region (45.32 to 46.86 Mbp) on chromosome X, 
SNPs (diamonds), scaffolds (green lines) and bovine ref sequences (black line)
           214 
Figure 7.18: Comparison of the FST values from the Romney versus Arapawa and the 
average FST value from the ovine HapMap study (Kijas et al., 2012a). The 0-1 line is 
plotted in red.         215 
Figure 7.19: Plot of the QQ slope against the selection intensity for each trait; weaning 
weight direct and maternal (?, ?), live weight at 8 months (?), fleece weight at 12 
months (+), dag at 3 and 8 months and breech bareness (?), wool bulk and length 
(?) and fibre traits (?)        217 
Figure 8.1: The first 4 principal components (PC) calculated from all animals for 
weaning weight. Romney (blue), Coopworth (green), Texel (yellow), Perendale 
(purple) and others/composites (grey)      242 
Figure 8.2: Manhattan plot of -log10(P) values of SNPs for weaning weight. Ordered on 
the ovine genome v2 map, P<0.0001 (solid red line), P<0.001 (red dash line) 251 
Figure 8.3: Manhattan plot of -log10(P) values of SNPs for fleece weight at 12 months. 
Ordered on the ovine genome v2 map, P<0.0001 (solid red line), P<0.001 (red dash 
line)          251 
Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
xviii 
 
Figure 8.4: Manhattan plot of -log10(P) values of SNPs for dag score at 3 months. 
Ordered on the ovine genome v2 map, P<0.0001 (solid red line), P<0.001 (red dash 
line)          252 
Figure 8.5: Manhattan plot of -log10(P) values of SNPs for dag score at 8 months. 
Ordered on the ovine genome v2 map, P<0.0001 (solid red line), P<0.001 (red dash 
line)          252 
Figure 8.6: Comparison of GWAS and GS -log10(P) values for weaning weight (A), 
weaning weight maternal (B), live weight at 8 months (C), fleece weight at 12 
months (D), dag score at 3 (E) and 8 (F) months. The 0-1 line also plotted (red line)
           253 
Figure 8.7: Comparison of dag score at 3 and 8 months (DAG3, DAG8) -log10(P) 
values from GS analysis. The 0-1 line also plotted (red line)   254 
Figure 8.8: QQ plot for weaning weight (A), fleece weight at 12 months (B), dag score 
at 3 months (C) and 8 months (D) ?log10(P) values. The 0-1 line is in black and the 
slope in red         255 
Figure 8.9: Comparison of the observed breed combined-accuracies (rC) versus the 
theoretical accuracy using Goddard?s equation for each trait, and for the breeds 
Romney (green), Coopworth (blue) and Perendale (purple). The line of equality is in 
grey          261 
 
List of abbreviations 
xix 
 
List of abbreviations 
AFEC  adult faecal egg count 
AFW  adult fleece weight 
BBREECH breech bareness 
BULK  wool bulk 
BV   breeding value 
chr   chromosome 
CURV  curvature 
CW  carcass weight 
DAG3  dag score at 3 months of age 
DAG8  dag score at 8 months of age 
dNTP  deoxynucleoside triphosphate  
EWT  adult ewe live weight 
FDCV  coefficient of variation of mean fibre diameter  
FDSD  standard deviation of mean fibre diameter 
FE   facial eczema 
FEC1  faecal egg count in summer 
FEC2  faecal egg count in autumn 
FW12  fleece weight at 12 months 
GBLUP  genomic best linear unbiased prediction 
GBV  genomic breeding value 
GC   genotype call score 
GWAS  genome-wide association study 
GS   genomic selection 
HapMap  haplotype map 
ISGC  International Sheep Genomic Consortium 
IWTO  International Wool Textile Organisation 
LENGTH  wool length 
Genetics of flystrike and dagginess in New Zealand dual-purpose sheep 
xx 
 
LFW  lamb fleece weight 
LW6  live weight at 6 months 
LW8  live weight at 8 months 
MAF  minor allele frequency 
MBV  molecular breeding value 
MED%  proportion of medullated fibres 
MFD  mean fibre diameter 
MT-EBV  multi trait-estimated breeding value 
NEM1  nematodirus egg counts in summer 
NEM2  nematodirus egg counts in autumn 
NLB  number of lambs born 
NZWTA  New Zealand Wool Testing Authority 
OFDA100  Optical Fibre Diameter Analyser 100 
OMIM  Online Mendelian Inheritance in Man 
PC   principal components 
PETA  People for the Ethical Treatment of Animals 
PEV  prediction error variance 
QQ  quantile-quantile 
RRS  reduced representational sequencing 
SAP  shrimp alkaline phosphate 
SIL   Sheep Improvement Limited 
ST-EBV  single trait-estimated breeding value 
SURV  lamb survival 
SURVm  survival maternal  
TBE  Tris base boric acid and ethylenediaminetetra-acetic acid buffer 
TBV  true breeding value 
WWT  weaning weight at 3 months 
WWTm  maternal weaning weight