Ordination of marker-trait association profiles from long- term international wheat trials Vivi...
-
Upload
conrad-garrison -
Category
Documents
-
view
213 -
download
0
Transcript of Ordination of marker-trait association profiles from long- term international wheat trials Vivi...
Ordination of marker-trait association profiles from long-
term international wheat trials
Vivi Arief, Pieter Kroonenberg, Ian Delacy, Mark Dieters, Jose Crossa and Kaye Basford
Outline
• Motivation
• Construction of the Wheat Phenome Atlas
• Three-way Principal Component Analysis of marker-trait association profiles
The Challenge
• To develop a detailed understanding of the heritable variation in the wheat genome
• To directly translate this knowledge into gains in wheat breeding
CIMMYT’s Wheat Program
• Targets wheat breeding in 6 agro-ecological regions around the world
• Contains a vast accumulation of knowledge, data and genetic resources
• Is publicly available
Wheat Trials
• 40 years of trials • With 10 or more international nurseries per
year• And 50 to 400 entries per nursery• At 50 or more locations around the world
→ 17m phenotypic data points on >80 traits across 13k lines evaluated in >10k field trials (data worth >US$500m)
Pedigree
• Pedigree information tracing the history of all 13k lines in the breeding programs
Seed Available
• Retained seed in storage from all trials available for low-cost high throughput genotyping (can give 26m marker data points)
Unique resource
• No other crop (to our knowledge) has this resource publicly available
Enabling Technologies
• ASREML (in the last 10 years)– An analysis program that can deal with the 17m
unbalanced data points
• ICIS (in the last 5 years)– Database to capture the pedigree, phenotype and
marker data
• DArTs (since 2006)– The first of high through-put affordable marker
systems– US$45 a DNA sample gives ~1,500 data points
Phenome Map
• Diagrammatic representation of the regions of a genome that influence heritable phenotypic variation for a trait
Phenome Atlas
• The integration of all phenome maps and a description of the methodologies that were used to produce the maps
Producing theWheat Phenome Atlas
• Focus on the ESWYTs
Elite Spring Wheat Yield Trials
Phenotypic dataAdvanced lines with high yield potential• 25 cycles from 1979/80 to 2004/05• 685 unique lines• 1445 trials across 400 locations• Phenotypic data for 20 traits
– 8 agronomic traits (including grain yield)– 3 rusts (leaf, stripe and stem)– 9 other foliar diseases
The Analysis
• Obtained BLUPs using ASREML• Fitted a separate residual and design for each trial
– ESWYT 1 to 13: RCB– ESWYT 14 to 25: -lattice
• Fitted separate models for combined association analysis and structured association analysis
– Combined association analysis: G model– Family structure: G model– Spatial structure: GGL model– Temporal structure: GGY model
Genotypic data
DArTs (Diversity Arrays Technology)
• Dominant markers 1; 0; X
• 1447 markers
• ~1.4 million data points
• 645 genotypes 599 unique genotypes (some are replicated)
The Analysis
Association analyses• Simple t-test
– for each trait– for each marker – for each structure
Marker order• ESWYT disequilibrium map
– No existing map shared more than 50% common markers with ESWYT
– Obtained using ESWYT dataset
Wheat Phenome Atlas Version 1.0
• Population structuresPedigree data Phenotypic dataMarker data
• Combination of population and environmental structures ESWYT cycle
• Environmental structures Mega-environmentPhenotypic data
10 Phenome maps:• Overall
x 2 analytical methods
A wheat phenome atlas
• Phenome map: dense QTL map for a trait• Phenome atlas: collection and description
of phenome maps
Increasing blue colour = increasing significance of positive association Increasing red colour = increasing significance of negative association
3 RustsStem (SR)Leaf (LR)Stripe (YR)
8 Agronomic traitsGrain Yield (GY)Kernel Size (KS)Plant Height (PH)Days to Heading (DH)Test Weight (TW)Grain Protein (GP)Lodging (LG)Shattering (SH)
9 Other foliar diseasesStripe rust on the spike (YS)Septoria tritici blotch (ST)Septoria nodurum blotch (SN)Spot blotch (SB)Powdery mildew (PM)Barley yellow dwarf (BYD)Fusarium leaf blotch (FN)Tan spot (TS)Xanthomonas (XT)
What did we find?
• Many Trait Associated Markers (TAMs) for a trait
• Multiple traits for a TAM• Association identified depend on:
– what germplasm included– where tested– when examined
What about the genotypes?
Phenome maps: markers traits
For selection: genotypes markers traits
Three-way principal component analysis
A TAM block
Markers in a linkage disequilibrium block showed significant association
log score 4
Association: positiveMarker : present
Association: negativeMarker : absent
Illustration 1
599 lines 288 TAM blocks 16 traits
599 lines 257 TAM blocks 15 traits
599 lines 213 TAM blocks 14 traits
Illustration 2
ME1 ME2
599 lines 202 TAM blocks 12 traits
599 lines 218 TAM blocks 17 traits
Three-way ordination:• Summarizes genotype TAM block trait
data• Reveals pattern in genotype TAM block
trait data• Parental selection• Genotype screening• Prediction of selection outcome
• Observed patterns depend on the genotype TAM block trait arrays used
Summary
Our Team
Vivi AriefIan DeLacyHailemichael DesmaeChristopher LambridesJacqueline BatleyDavid EdwardsMark DietersIan GodwinKaye Basford
Jose CrossaSusanne DreisigackerTom PayneRavi SinghEtienne DuveillerGuy DavenportYann ManesMarilyn WarburtonGraham McLarenHans-Joachim BraunJonathan CrouchRodomiro Ortiz
Peter WenzelEric HuttnerAndrzej Kilian
Also