MAPNet
2014
Abstracts
Day 1 - Tuesday, 21 October
Session 1 – Chair: Suzanne Rowe
Reducing uncertainty with noisy data: Challenges and opportunities of using sequence data to
genetically improve livestock and crops
Hans D. Daetwyler, Arjan Tolkamp, Bolormaa Sunduimijid, Amanda J. Chamberlain, and Ben J.
Hayes
Genomic selection coupled with reliable markers from medium density single nucleotide
polymorphism (SNP) chips has been responsible for an ongoing and successful transformation of
animal and plant breeding in the past decade. Increasing the number of markers may enable further
gains in prediction accuracy. The ultimate marker density is whole-genome sequence and the 1000
Bull Genomes Project has collected sequence from a large number of cattle to call variants that are
distributed to partners. It constitutes the most complete inventory of SNP and indels in cattle and
forms the basis for a wide variety of analyses, from genome-wide association and selection to
identification of causative mutations. One key challenge of sequence data is that it contains
erroneously called bases in reads that may translate into false variants, which, in turn, cause issues
during imputation. Filtering of raw sequence and called variants is a critical step to increase the
chance of success for analyses. Validation of such filters is required for tuning, which will be
demonstrated with examples. Finally, some early success stories of using whole-genome sequence
data will be presented.
Ovine 50k to HD Imputation
Matt Bixley, Rudi Browning, Ken Dodds, Suzanne-Rowe, Hayley Baird, Dianne Hyndman, Rayna
Anderson, Shannon Clarke, Benoit Auvray
Imputation from the Ovine 50k SNP chip to the Ovine HD Chip is an accurate and valuable tool for
Genomic Selection and GWAS. 50K to HD chip accuracies of 0.98 were routinely achieved with
suitable reference populations and no use of pedigree. Romney, Perendale and a Composite breed
were used to investigate imputation accuracy both within and across breed.
Preliminary results of genomic prediction for residual feed intake in multibreed and crossbred
beef cattle in Canada
Duc Lu1,2, John Crowley1,3, John Basarab1,4, Changxi Li1,4, Zhiquan Wang1, Mary DePauw1, Graham
Plastow1, Steve Moore1,5, Steve Miller1,2,6, Paul Stothard1
1Livestock Gentec, University of Alberta, Edmonton, AB, Canada; 2AgResearch, Invermay, Mosgiel,
Otago, New Zealand; 3Canadian Beef Breeds Council, Calgary, AB, Canada; 4Agriculture & Agri-Food
Canada, Canada; 5Queensland Alliance for Agriculture & Food Innovation, University of Queensland,
QLD, Australia; 6Centre for Genetic Improvement of Livestock, University of Guelph, ON, Canada
The key deliverable of this activity was to develop molecular breeding values (MBV) for residual feed
intake (RFI) based on high density genotypes. There were 6796 animals (Kinsella Ranch; University
of Guelph; Phenomic Gap Project; NSERC Hereford) with RFI data and genotypes (Illumina
Bovine50SNP Chip or Affymetrix HD Bovine Chip) being used in the analysis. MBV was developed
using BLUP and Bayesian approach with highest accuracies of 0.62 and 0.54, respectively, when the
50K genotypes was used. Initial analysis using the HD genotypes did not show improvement in
accuracy. However further edits are underway for analysis with the HD genotypes.
Development of breeding infrastructure for the NZ Sheep industry: an example from FarmIQ
Michael Lee
Abstract not submitted as yet
Session 2 – Chair: Peter Amer
Application of genetics and genomics to New Zealand aquaculture
JE Symondsa, MD Camarab, SP Walkera, S Clarkec and J McEwanc
aNational Institute of Water and Atmospheric Research (NIWA), Ruakaka, New Zealand
bCawthron Institute, Nelson, New Zealand
cAgResearch, Invermay, Dunedin
Aquaculture is a fast growing sector of agricultural production in New Zealand and farmed
GreenshellTM mussels (Perna calaniculus), Pacific oysters (Crassostrea gigas), king salmon
(Oncorhynchus tshawytscha) and paua (Haliotis iris) currently produce over $400 million in revenue
(aquaculture.org.nz). The industry has plans to expand production of these species plus add
emerging high-value species such as kingfish (Seriola lalandi) and hāpuku (Polyprion oxygeneios) to
this list. Over the last two decades improvements to aquaculture stock performance using genetic
technologies have become established in New Zealand. The main approach is family pedigree based
selection with the focus on the production and evaluation of families. Defining and measuring the
traits to be selected, and their relative economic value, continues to evolve as the industry develops
and new challenges arise. All programmes have verified the significant potential for genetic
improvement and the moderate to high heritabilities for many of the production traits evaluated are
encouraging. Genomics tools are also being developed and the future will likely see a combination of
traditional and genomic selection approaches for some species.
GreenshellTM mussel genomics
Rachael Ashby1,3, Shannon Clarke2, Neil Gemmell3, Chris Brown1
Biochemistry, University of Otago2. AgResearch, Invermay, 3. Anatomy, University of Otago.
The GreenshellTM Mussel (Perna canaliculus) is a bivalve mollusc native to New Zealand. It is of large
economic importance to the New Zealand Aquaculture industry. Techniques are currently being
developed to selectively breed mussels for desirable traits to improve mussel farming and increase
export value. Current sequencing technologies have enabled the hitherto intractable problems to be
approached by sequence analysis. In mussel we are currently undertaking de novo assembly and
annotation of both the genome and transcriptome. These will be used to identify genetic markers and
genes of interest to biology and/or aquaculture.
Genomic selection in Atlantic Salmon (Salmo salar)
John McEwan, Rayna Anderson, Jónas Jónasson, Theódór Kristjánsson, Eduardo Rodriguez, Ólafur
Kristjánsson, and Shannon Clarke
Stofnfiskur (www.stofnfiskur.is) produces Atlantic Salmon eggs for sale, year round. Egg production is
the product of a closed intensive breeding scheme. All brood stock are mated, hatched and reared in
temperature, light and salinity controlled environment using sea- and freshwater is pumped from
geothermal- and freshwater wells ensuring the best disease free environment in the world. Parentage
and identity in brood stock are tracked via use of RFID tags. Eggs for commercial production are
shipped to rearing farms in Norway, Faroe Islands, Scotland, Ireland and Chile. The volume of salmon
product produced from this breeding programme is comparable to the New Zealand lamb meat
production.
Currently, disease and meat quality traits selection in the broodstock is based on full-sib family BLUP
estimates via progeny tests, typically using DNA parentage for the latter. Given that up to 4 disease
traits as well as meat quality are involved, this approach is expensive and does not allow selection
within full-sib families for these traits.
An alternative approach is to use genomic selection and a structured approach to implement this
using the latest technology is being undertaken. The work includes: genome sequencing and
assembly including use of Illumina Long Read Technology, SNP discovery and physical and linkage
mapping of scaffolds using genotyping by sequencing (GBS), followed by use of existing progeny test
DNA as a training set to implement genomic selection. Progeny test and selection candidates will be
genotyped at a lower density either via GBS or an equivalent technology. This approach is part of an
exemplar project to demonstrate low GBS can be implemented at low cost into non-traditional
species.
Genotype by environment interactions in forest genetics
Yongjun Li
Scion (New Zealand Forest Research Institute), P.O. Box 3020, Rotorua 3046, New Zealand
The phenomenon whereby one genotype that is superior in one environment might be inferior in
another environment is called genotype by environment (GE) interactions. Identifying GxE
interactions is important to maximize genetic improvement of forestry. It enables the best suitable
genotypes to be selected for a specific environment, which maximizes genetic gain in that
environment. It also enables the genotypes that are not sensitive to specific environment to be
selected so that breeders can expect stable performance over all environments. GxE level in forest
genetics is measured using genetic correlation of the same trait between environments. There are
three main methods to estimate GxE in forest genetic: Type B genetic correlation, factor analytic
model and Biplot analysis. Type B genetic correlation method is suitable to a situation where there is
good connectivity among environments and the number of environments is not large. Factor analytic
model can accommodate large number of trials and poor connectivity between trials. GGE Biplot
analysis is suitable for identifying genotypes for deployment. Implication of GE interactions in forest
breeding and deployment are also discussed.
Wednesday, 22 October - Day 2
Session 1 – Chair: Gemma Jenkins
Consequences of selection in both domestic sheep and populations of wild bighorn
James Kijas, Agriculture Flagship, CSIRO
The patterns of genetic diversity present within animal genomes can be used to investigate aspects of
population history, evolution and the consequences of adaptation. This study sought to use patterns
of diversity to identify selective sweeps that have occurred during the evolution of ovine populations
with contrasting population histories. Three datasets were collected and analysed. First, genome
sequencing of DNA pools was used to identify 3.23 M SNP from three North American bighorn
populations. Second, 68 domestic sheep from a 43 breeds were individually sequenced to identity
over 20 M SNP. Finally, the SNP50 Beadchip was used to genotype individuals from a number of
Merino selection lines. In each case, analysis sought to identify genomic regions that exhibit an
increase in homozygosity consistent with the action of selection. Selective sweep regions
wereidentified from each dataset, along with the genes contained with them. Genes controlling
pigmentation, body conformation and production traits were identified within domestic sheep, while
genes with known roles in horn morphology and early growth were identified in bighorn. The results
extend our understanding of genes likely to have undergone positive selection during the evolution of
sheep and more broadly point to the molecular consequences of adaptation.
Complex variation in measures of general intelligence and cognitive change
Suzanne J Rowe,1 Gail Davies,2 Sarah E. Harris,2,3 David J. Porteous,2,3 David C. Liewald,2 Geraldine
McNeill,4 John M. Starr,2,5 Ian J. Deary,2,6,8 Albert Tenesa1,7
1The
Roslin Institute, Royal (Dick) School of Veterinary Studies, The University of Edinburgh, Easter Bush
Campus, Midlothian, EH25 9RG. Scotland. UK, 2Centre for Cognitive Ageing and Cognitive Epidemiology,
University of Edinburgh, 7 George Square, Edinburgh EH8 9JZ, UK, 3Medical Genetics Section, The University of
Edinburgh Molecular Medicine Centre, Institute of Genetics and Molecular Medicine, Western General Hospital
Edinburgh, UK, 4Institute of Applied Health Sciences, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD,
5Alzheimer Scotland Dementia Research Centre, The University of Edinburgh, 7 George Square, Edinburgh, UK,
6Department of Psychology, University of Edinburgh, 7 George Square, Edinburgh EH8 9JZ, UK, 7MRC HGU at
the MRC IGMM, University of Edinburgh, Western General Hospital, Crewe Road South, Edinburgh. EH4 2XU.
UK
One approach to bridge the gap between estimates of global genetic variation attributed to a trait, and
that derived from point estimates of fixed marker effects is the recently proposed “regional heritability
mapping”. Combining information from multiple SNPs may capture a greater amount of genetic
variation than from the sum of individual SNP effects. Regions may capture variation from multiple
common variants of small effect, multiple rare variants or a combination of both. We describe regional
heritability mapping of human cognition. Measures of crystallised (g c) and fluid intelligence (gf) in late
adulthood (64-79 years) were available for 1806 individuals genotyped for 549,692 autosomal single
nucleotide polymorphisms (SNPs). The same individuals were tested at age 11, enabling us the rare
opportunity to measure cognitive change across most of their lifespan. We divide 547,750 SNPs
ranked by position into 10, 908 overlapping regions of 101 SNPs and estimate the genetic variance
explained, an approach that resembles classical linkage methods. We also estimate the genetic
variation explained by individual autosomes and by SNPs within genes. We estimate empirical
significance thresholds separately for each trait from whole genome scans of 500 permutated data
sets. The 5% significance threshold for the likelihood ratio test of a single region ranged from 17-17.5
for the three traits. This is the equivalent to nominal significance under the expectation of a chisquared distribution (between 1df and 0) of P < 1.44 E-05. These thresholds indicate that the
distribution of the likelihood ratio test from this type of variance component analysis should always be
estimated empirically. Furthermore, our analyses show that estimates of variation explained by these
regions can be grossly overestimated. After applying permutation thresholds, a region for g f on
chromosome 5 spanning the CTXN and SLC2A2 genes, previously identified in candidate studies for
other cognition-related diseases, is genome wide significant at a 10% threshold, providing a target
region for high throughput sequencing of these uniquely informative cohorts.
Transcriptome profiling of the abomasal lymph node of Scottish Blackface lambs with
divergent phenotypes for resistance to gastrointestinal nematodes
Kathryn M. McRae1, 2, Barbara Good3, James P. Hanrahan3, Mary J. O’Connell2& Orla M. Keane1
1. Animal & Bioscience Department, Teagasc, Grange, Dunsany, Co. Meath, Ireland
2. Bioinformatics & Molecular Evolution Group, School of Biotechnology, Dublin City University,
Glasnevin, Dublin 9, Ireland
3. Animal & Bioscience Department, Teagasc, Athenry, Co. Galway, Ireland
In Ireland, two sheep breeds have been extensively studied with respect to gastrointestinal nematode
(GIN) resistance, and it has been demonstrated that the Texel breed is more resistant to nematode
infection than the Suffolk breed. However to date no studies have examined GIN resistance among
Scottish Blackface sheep in Ireland. The objective of this study was to identify Scottish Blackface
lambs that differed in their resistance to GIN infection and identify genes and biological processes
important for the difference in resistance status.
An animal selection model was developed that reliably identified Scottish Blackface lambs that
differed in resistance to GIN, as measured by faecal egg count (FEC). Selected resistant (low FEC, n
= 10) and susceptible (high FEC, n = 10) animals were given a controlled challenge of 30,000
Teladorsagia circumcincta L3 larvae. In order to identify genes and biological processes associated
with the acute response to GIN infection in resistant and susceptible individuals, the abomasal lymph
node transcriptome was examined at 7 and 14 days post infection.
At seven days post infection resistant animals appear to be generating a more effective immune
response, as biological processes involving leukocyte migration, the inflammatory response and
migration of antigen presenting cells were upregulated, whereas in susceptible animals this response
is delayed until ~14 days post infection. The earlier immune response generated by the resistant
animals may therefore be influencing the larval stages of T. circumcincta, which could result in
shorter, less fecund adults.
Session 2 – Chair: Tony Merriman
Integration of population-scale DNA and RNA sequence data for the identification of causative
genes in Bos taurus
Dr Matt Littlejohn1,2, T Law2, K Tiplady1, R Sherlock1, M Keehan1, T Lopdell1, C Harland1, T Johnson1,
A Scott1, V Obolonkin1, Wouter Coppieters3, Michel Georges3, K Lehnert1,2,R Spelman1, S Davis1, R
Snell2
1Livestock Improvement Corporation, Hamilton, New Zealand; 2University of Auckland, Auckland,
New Zealand; 3University of Liege, Liege, Belgium.
Genome-wide association studies have identified numerous quantitative trait loci (QTL) underlying
physiological traits in humans, mice, cattle, and other species, though the identification of causative
genes and variants at these loci remains challenging. We describe two complementary approaches
being applied to identify causative genes and variants using large, multidimensional ‘omics’ datasets
in Bos taurus.
The first method exploits large scale transcriptomic data to identify candidate causative genes through
alignment of gene expression QTL (eQTL) and QTL for physiological traits. We will present the utility
of this approach in cattle, describing the integration of large scale milk composition GWAS data
(N=66,262 cows) with high-depth mammary RNA sequencing data (N=406 cows).
The second approach aims to use genome sequencing and imputation to map causative variants
directly. We have sequenced the genomes of 556 mixed ancestry cattle to a total depth of >6000x,
and will report use of these data to impute sequence-resolution genotypes into a commercial dairy
cow population (N=66,262). GWAS results using these data will be presented, highlighting a new
locus with major impacts on milk composition.
The potato genome- from a haploid reference to autotetraploid reality
Jeanne Jacobs
The New Zealand Institute for Plant & Food Research Limited, Private Bag 4704, Christchurch 8140,
New Zealand
The reference genome sequence of potato, developed by the Potato Genome Sequencing
Consortium (www.potatogenome.net), is a great resource for expanding the knowledge on the
gene(s) underpinning traits of interest. The first draft of the potato genome together with RNAseq data
has facilitated the construction of a 8.3k potato SNP chip (SolCAP; http://solcap.msu.edu). To
genotype PFR breeding lines, cultivars, and genetic mapping populations we have used this SNP chip
and have re-sequenced five potato cultivars/ breeding lines, including the parents of three biparental
mapping populations. While the reference genome sequence was developed using a homozygous
diploid potato (DM1-3 516 R44) with only one haplotype, cultivated potato is autotetraploid and highly
heterozygous. The autotetraploid nature of potato complicates the analyses of the genotypic data,
generated via SNP chip analysis and resequencing, as up to four haplotypes can be observed. This
presentation will address some of the challenges associated with the move from a reference genome
sequence based on one haplotype to using highly heterozygous tetraploid material for studying the
genetic variation across the genome.
Application of genome-assisted breeding tools in fruit crops
Satish Kumar
The New Zealand Institute for Plant & Food Research Limited, Hawkes Bay Research Centre,
Havelock North 4130, New Zealand
Understanding the genetic basis of various fruit quality traits and physiological disorders is important
for the development of new apple cultivars. Availability of apple genome sequence and cost-effective
high throughput genotyping platforms have helped develop genome-based crop improvement
methods such as genome-wide association (GWA) mapping and genomic selection (GS). Using an
8K Apple SNP Array we compared patterns of linkage disequilibrium between a breeding population
and a collection of germplasm accessions. GWA mapping of economically important traits validated
the role of previously identified candidate genes, and also revealed new genomic hotpots. For most
traits, the distributions of SNP effects were consistent with a polygenic quantitative model suggesting
that genomic selection (GS) would be a better tool for selection purposes. Random-regression best
linear unbiased prediction (RR-BLUP) and the Bayesian LASSO method were used to obtain genomic
breeding value (GEBV), and compared using a cross-validation approach for their accuracy to predict
unobserved BLUP-BV. The future of fruit breeding could involve genotyping-by-sequencing (GBS) to
compare individual plant genomes, allele mining, and improving our understanding of genotypephenotype relationships. Results from GS and GWA studies, and opportunities and challenges using
GBS in fruit breeding will be presented.
An integrated framework to identify breakpoints of events inside a large copy-number variable
region
Hoang T Nguyen
1,2,3
1,3
, Tony R Merriman , Michael A Black
1,3
1
Departments of Biochemistry, 2Mathematics and Statistics, University of Otago, Box 56, Dunedin,
3
New Zealand, Virtual Institute of Statistical Genetics, New Zealand
Copy-number variation (CNV) has been associated with increased risk of complex diseases, however,
associations are not widely replicated. One possible reason is the technical difficulty in accurately
measuring copy number. Another reason can be that precise breakpoints of some copy-number
regions (CNVRs) have not been identified. Pathways of disease can be different if breakpoints occur
at different positions (e.g., some breakpoints can disrupt functional regions of genes). Here, we report
an integrated work flow, which combines read-depth and split-read approaches using high-throughput
sequencing (HTS) data, to identify breakpoints of different events inside a large CNV region. The
main approach of the pipeline is that it uses information from multiple samples to obtain breakpoints.
To calculate the performance, the pipeline and three other HTS-based packages were applied to
specific genomic loci from simulated and real data sets. The pipeline showed higher true positive
rates than the other applications, suggesting that this pipeline can be used to exactly pinpoint
breakpoints at specific loci using HTS data.
A tale of two genomes - de novo next generation genome sequencing of white clover and its
progenitors
Andrew Griffiths1,2, Roger Moraga1, Stig Andersen3, Anar Khan
1AgResearch, Grasslands Research Centre, Private Bag 11008, Palmerston North 4442, 2Pastoral
Genomics, ℅AgResearch, Grasslands Research Centre3Aarhus University, Gustav Wieds Vej 10,
Aarhus, Denmark DK-8000
White clover (Trifolium repens L.) is a principal legume in temperate pastoral systems, and access to
a reference genome provides a platform to understand and improve the species. Polyploidy, a
common feature of many plant crops, impedes assembly when developing de novo reference
genomes based on short-read sequencing platforms. This impediment is exacerbated in white clover
as it is an out-breeding allopolypoid derived from two closely-related progenitors. The resulting high
levels of intra-homoeologous sequence diversity combined with the presence of genetically similar
homoeologues (progenitor genomes) provide a “challenge” when designing a strategy for sequencing
white clover. As part of the Pastoral Genomics Consortium in collaboration with Aarhus University
(Denmark), we are close to completion of the first annotated draft genome of white clover and its
progenitors. The latest results will be presented providing insight into the genomes and how it
improves our understanding of evolution of a species and its application to forage improvement.
Session 3 – Chair: Phil Wilcox
New genetic approaches reveal rapid evolution in New Zealand species
Jonathon Waters
Abstract not submitted yet
Conditional analysis of 30 serum urate loci identifies 25 additional independent effects
Tony Merriman
Background/Purpose: Single variants in 30 genetic loci have been associated with serum urate levels
in Europeans by meta-analysis of summary statistics of 48 individual genome-wide association study
(GWAS) data sets. Identifying independent effects can help fine map causal genes and reveal
functional mechanisms of genetic variation, and has been successful in other complex phenotypes.
Our aim was to test for the presence of independent effects at the 30 urate loci.
Methods: Summary level statistics from the Kottgen et al. GWAS were used in collaboration with the
Global Urate Genetics Consortium. The Genome-Wide Complex Trait Analysis (GCTA) package was
used to test for association conditional on the lead single nucleotide polymorphism (SNP). A total of
9713 HapMap2 genome-wide imputed genotypes from European participants of the Atherosclerosis
Risk in Communities study were used as a reference. An independent effect at each locus was
defined as an association signal, after conditional analysis, of P<0.05 divided by the number of SNPs
analyzed. Further rounds of analysis were conducted if SNPs remained significant after conditioning.
The percent variance explained for each SNP was calculated by the formula ^2*(var(X) / var(Y)).
Results: Twenty-five additional independent effects were detected at 14 of the 30 urate loci (Table).
The percent variance explained by the 30 lead SNPs was calculated to be 4.83%, with the additional
SNPs explaining a further 0.81% of variance (total 5.64%).
Conclusion: The independent effects provide evidence for multiple etiological variants at the serum
urate loci in Europeans, emphasizing the complex genetic control of serum urate levels. These results
are an important initial step in fine-mapping the causal variants. For example, one of the independent
effects at SLC2A9 (rs3775948) was the lead SNP in a urate GWAS in East Asian individuals,
indicating both shared and unique genetic effects between East Asians and Europeans. The use of
multiple ancestral groups will be important in fine-mapping.
Apoptotic pathway regulation by Pitx1 and Tbx4 may unravel the etiology of Clubfoot
Ng, J1, Rios, J2. and Wilson M.J1.
1Department
2Texas
of Anatomy, University of Otago, Dunedin, New Zealand
Rite Scottish Hospital for Children, Dallas, Texas, USA.
Clubfoot, otherwise known as talpes equino varus is a common congenital deformity where its
etiology is not well understood. Clubfoot symptoms consist of a unilateral or bilateral inward turning of
the foot and the heel pointing downwards. Clubfoot has an incidence rate of one in 1000 children and
a higher incidence in the Maori population with seven in 1000. Evidence has shown that 20% of
clubfoot cases are non-idiopathic and is associated with other congenital diseases such as distal
arthrogryposis, myelomeningocele. Genetic syndromes include trisomy 18 and chromosome 22q11
deletion syndrome which has a higher incidence rate. However, the majority of clubfoot cases are
idiopathic. Patient studies have shown microduplication of chromosome 17q23 which includes T-box
4 (TBX4). Additionally mutation in paired-like homeodomain 1 (PITX1) is linked to clubfoot, and
therefore shows the importance of normal expression of these genes is vital during embryonic
development. These two transcription factors are expressed only in lower limb during development,
further supporting the notion these two genes are likely involved in idiopathic clubfoot.
Preliminarily data carried out by Dr Wilson using Chromatin Immunioprecipation followed by
sequencing (ChIP-seq) on Pitx1 and Tbx4 in 12.5 days post coitum (dpc) mouse limb budshas shown
a strong enrichment of P53 apoptotic pathway genes. Other enriched genes include insulin growth
factor binding protein (Igfbp3) and apoptotic-associated genes such as capase (Casp) 8 and
Casp8ap2. We hypothesised that this strong enrichment from Tbx4 and Pitx1 ChIPtargeting apoptotic
genes could be one of the mechanisms that are important for the development of the lower limb and
abnormality in this expression could lead to a clubfoot phenotype.
Using the preliminary ChIP-seq data, we will confirm the enrichment of the apoptotic genes targets
between the forelimb and hindlimb in mouse embryos at 12.5 dpc. Quantitative real time PCR will be
used quantify the relative expression of candidate genes between forelimb and hindlimb buds. Whole
mount in situ hybridisation will allow visualisation of the mRNA expression of the candidate genes
throughout the embryo. Lastly, to further validate the importance of these candidate genes within the
apoptotic pathway, we will overexpress Pitx1 and Tbx4 in cell cultures, which will determine whether
these genes cause an increase or decrease of apoptosis within the cell culture. We aim to understand
the association between the Pitx1-Tbx4 expression and the apoptotic importance in the etiology of
clubfoot and normal lower limb development.
Possibilities of the use of genotyping to reduce bias in genetic evaluation
John Holmes
An important question in genetic evaluation is breeding value comparability. This is particularly true
when comparing animals across fixed groups, such as flock or contemporary group.
Bias can be introduced into genetic evaluation if the fixed groups are poorly connected.
Connectedness is defined as the genetic links joining one fixed group to another.
The effect of lack of genetic connectedness amongst flocks evaluated in the NZ sheep industry is not
well understood. A lack of connectedness will probably lead to a reduction in the rate of genetic gain
achieved.
This project aims to understand connectedness and its consequences in the NZ sheep industry.
In particular, it will re-evaluate various connectedness measures and investigate the use of genotype
data as a method to recover true connectedness missed in traditional pedigree based methods.
Thursday, 23 October - Day 3
Session 1 – Chair: Shannon Clarke
Reproducibility and Mendelian Inheritance of SNPs using Exome Capture-based Genotypingby-Sequencing in Pinus radiata D.Don
P.L. Wilcox1, E.J. Telfer1, Y. Li1, N. Graham1, J. Butcher2, J. Hay2, M. Resende3, L. Neves3 and M.
Kirst3
1Scion (New Zealand Forest Research Institute Ltd), Private Bag 3020, Rotorua, NZ
2Radiata Pine Breeding Company, PO Box 1127, Rotorua, NZ
3Rapid Genomics LLC, 747 SW 2nd Avenue, Gainesville, Florida 32601, USA
We are evaluating exome-based genotyping-by-sequencing for the implementation of genomic
selection in the New Zealand Pinus radiata breeding programme. Using SNP data generated from an
80K exome-capture probe panel on 24 diploid and 6 haploid samples, we examined reproducibility
across technical and biological replicates as well as Mendelian inheritance. Of the 80K probes, 51
903 probes were found to be polymorphic, yielding 284 000 SNPs. Filtering SNPs with <10 average
read depth across all samples reduced this to 49 701 SNPs from 18 222 probes. Read depth varied
across DNA samples and was strongly correlated with the proportion of missing data. Reproducibility
in duplicate samples was 93.3% for >25X average read depth in the filtered data set, but decreased to
79.3% with lower read depths (<15X) in the unfiltered data set. Mendelian inheritance was examined
using 8 diploid trios and 2 parent-megagametophyte pairs. In the filtered data set, the inheritance of
36 793 SNPs matched expected inheritance among at least 7 of 10 trios and parentmegagametophyte pairs, while 3890 SNPs were completely consistent with expected inheritance
among all trios and parent-megagametophyte pairs. The proportion of homozygous loci per sample
was also correlated with read depth, with samples exceeding 20X read depth exhibiting the expected
number of heterozygous SNPs based on average minor allele frequency, while samples with lower
read depth had an excess of homozygous SNPs. Our results suggest (a) per sample read depth of
20X is required to achieve minimum accuracy, and (b) filtering criteria should be applied to individual
genotypic data points rather than to each locus based on average read depth across samples.
Expansion of the global repertoire of pea powdery mildew resistance genes
G.T.M. Poulter, P.A. Stockwell, R. Lee, R.T.M. Poulter and R.C. Macknight
A significant amount of the annual pea (Pisum sativum) crop is lost because of powdery mildew
infection (Erysiphe pisi). Chemical control methods are expensive and often environmentally harmful.
Establishing varieties of pea resistant to powdery mildew would be of great benefit. A near exhaustive
search of pea and pea-related germplasm has yielded a limited number of powdery mildew resistance
genes. One of these genes, ErB, is found in a wild relative of pea - Lathyrus belinensis.
Discovering the genetic sequence of ErB could allow the rapid introduction of powdery mildew
resistance into cultivated pea. The resistant phenotype was introduced into a susceptible Lathyrus
background through inter-specific hybridisation. This hybrid was then used to generate a nearisogenic line (NIL) through 10 successive backcrosses to the susceptible parent. Selection for the
resistant phenotype was made at each generation.
Analysis of the transcriptomes of the wild Lathyrus belinensis, of the susceptible Lathyrus, and of the
near-isogenic line (NIL) carrying ErB enabled mapping of the introgressed region that confers
resistance. This target region includes many NBS-LRR class resistance genes, and is confidently
inferred through comparison to available legume genome sequences (Medicago truncatula/Barrel
medic and Cicer arietinum/Chickpea).
Candidate gene sequences for ErB were identified in this target region through bioinformatic analysis.
Sequences encoding NBS-LRR genes made up the primary candidate set. To refine the number of
candidate genes, we aim to identify NILs carrying a recombination event within the target region. This
will allow us to fine-map the target region and identify genes that co-segregate with the resistant
phenotype. A detailed high-resolution melt (HRM) based assay was designed to screen for the
desired recombinants. HRM provides a highly sensitive, cost-effective and high-throughput system for
screening a large number of individuals. The assay provides SNP genotyping at 5 loci; these SNPs
being first identified through the previous Lathyrus complete transcriptome sequencing.
Exploratory experiments have been carried out to inform the design of future work. This includes the
successful transient expression of genes in pea, allowing for candidate genes to be directly assayed
in planta.
Genotyping-by-Sequencing (GBS): Efficiency and reproducibility checks
Tracey van Stijn, Rudiger Brauning, Ken Dodds, John McEwan and Shannon Clarke
Genotyping-by-Sequencing (GBS) has the potential to be a cost effective, reproducible and highthroughput SNP genotyping method. We have been investigating GBS in a number of livestock
species, with an emphasis on Sheep. To establish an accurate and reproducible method we have
been evaluating each step. In addition, SNPs located in the GBS targeted genome regions were
included on the sheep High Density (HD) Illumina SNP chip to allow forward genotyping capability.
We have compared genotype calls between the two genotyping methods as well as whole genome
sequencing (between 10x – 20x genome coverage).
Molecular markers – Genetic diversity, germplasm rationalisation and maps
Kioumars Ghamkhar
Molecular markers play an important role in plant pre-breeding research and are already being used
in plant breeding. Searching for variation in germplasm or populations is the basis of this paper.
Molecular marker transferability across species to exploit elite molecular markers for cultivar
identification using linkage maps, correlations between non-molecular and molecular data, exploring
genetic diversity for representative lines/accessions and core collections will be presented.
Identification of areas of unknown diversity (germplasm diversity gaps) and maximum diversity
(germplasm diversity hotspots) for targeted and rational seed collection in future will be also
described.