The coalition between Italian goats and
Italian researchers: the Italian Goat
Consortium
Paolo Ajmone Marsan and the Italian Goat Consortium
Institute of Zootechnics
Università Cattolica del Sacro Cuore
Piacenza, Italy
[email protected]
Cardiff, 17/06/2014
Outline
IMMAGINE
IMMAGINE
• The Italian Goat Consortium
• SNP diversity in Italian goats
• Perspectives
Italian goat Consortium
• Hard time for economy in Europe
– Harder in Italy than in Central/Northern Europe
• Even harder for research funding
– Very difficuly for small ruminant research
» No way for goat diversity!
• From crysis to opportunity
– Modest seed funding of a project started in
2008(Innovagen funded by the Ministry of Agriculture)
• Coalization and Coordination…..
– Definetely a new model for Italian scientists…….
Italian goat Consortium
Paola Crepaldi, Coordinator
Università degli Studi di Milano, ITALY
Fabio PILLA, Maria Silvia D'ANDREA
Università degli Studi del Molise, Campobasso, ITALY
Paolo AJMONE-MARSAN, Nicola BACCIU, Lorenzo BOMBA,
Licia COLLI, Marco MILANESI
Università Cattolica del Sacro Cuore, Piacenza, ITALY
Antonello CARTA, Tiziana SECHI
AGRIS, Loc. Bonassai, Sassari, ITALY
Stefania CHESSA, Bianca CASTIGLIONI
Consiglio Nazionale delle Ricerche, Lodi, ITALY
Donata MARLETTA, Salvatore BORDONARO
Università degli Studi di Catania, ITALY
Salvatore MURRU
Associazione Nazionale della Pastorizia, Roma, ITALY
Riccardo NEGRINI, Raffaele MAZZA
Associazione Italiana Allevatori, Roma, ITALY
Giulio PAGNACCO, Beatrice COIZET, Letizia NICOLOSO,
Università degli Studi di Milano, ITALY
Alessio VALENTINI
Università degli Studi della Tuscia, Viterbo, ITALY
Paola Crepaldi
[email protected]
Pooling local efforts and
resources for the genomic
characterisation of Italian
goat breeds
www.italiangoatconsortium.eu
Sampling
Camosciata delle Alpi
Saanen
Orobica
Bionda dell’Adamello
Val Passiria
Valdostana
Teramana
Grigia Ciociara
Nicastrese
Sarda
Aspromontana
Maltese
Girgentana
Argentata dell’Etna
50K Illumina goat SNP chip
• Discovery on 6 breeds (meat, mixed and milk)
• Detection of ~12 million variations with > 10 millionSNPs
• 60,000 SNPs (spaced on the genome, with >0.2 MAF,
>0.8 Illumina ADT score…)
• 52,295 successful loci (tested with 288 goat DNA
samples from 10 different breeds)
• Pseudochromosomes aligned on cattle
• Details on www.goatgenome.org
• Sequencing and novel de novo assembly on going at
USDA
Dataset cleaning
• Filtering exclusion threshold
– MAF < 1%
– Missing (SNP) > 5%
– Missing (animal) > 5%
– HW within breed FDR > 20%
• Working Dataset
– 15 breeds
– 350 animals (15-32 per breed)
– 51,136 SNPs
SNPchip affected by ascertainment bias (EU Nextgen
project) but highly informative for the Italian gene pool
Within breed MAF
distribution
CAM
SAA
SAR
ARG
BIO
X=0
VPS
0<X≤0.05
ASP
0.05<X≤0.1
NIC
0.1<X≤0.2
GCI
0.2<X≤0.3
VAL
0.3<X≤0.4
MAL
0.4<X≤0.5
ORO
SAM
GIR
TER
0
4000
8000 12000 16000 20000 24000 28000 32000 36000 40000 44000 48000 52000
Expected vs Observed
Heterozygosity
K=2
Alps
Center
South & Islands
Cross Validation error
plot
ADMIXTURE Software (10 runs)
The Best K (K=11)
Alps
Center
South & Islands
Geographic
distribution of
11 genomic
components
Neighbour Net based on
Reynolds distance
Reynolds Distance
DRe ynolds =
å (x - y )
1- å x y
i
1
2
i
i
i
i
i
2
Principal Component
Analysis
CAM = Camosciata (Alpine)
VAL = Valdostana
Teramana
SAA = Saanen
BIO = Bionda
ORO = Orobica
VPS = Valpassiria
GCI = Grigia Ciociara
ARG = Argentata
NIC = Nicastrese
ASP = Aspromontana
Alps
Center-South
MAL = Maltese Siciliana
GIR = Girgentana
TER = Teramana
SAM = Maltese Sarda
Maltese
SAR = Sarda
Principal Component
Analysis
CAM = Camosciata (Alpine)
West-East Alps
VAL = Valdostana
SAA = Saanen
BIO = Bionda
ORO = Orobica
Aspromontana
VPS = Valpassiria
GCI = Grigia Ciociara
Girgentana
ARG = Argentata
Maltese
NIC = Nicastrese
ASP = Aspromontana
MAL = Maltese Siciliana
GIR = Girgentana
TER = Teramana
Orobica
SAM = Maltese Sarda
SAR = Sarda
r2
LD in Chromosome 6
Distance (Mb)
Ne
Historical Ne of Italian goat breeds was estimated using SNeP*. For each pair of
SNPs within a chromosome the LD is calculated according to Hill & Robertson
(1968) using the method of Sved (1971) and correcting for sample size and
mutations (Weir & Hill 1980, Hayes 2003, Corbin et al. 2012).
5000
4500
ARG
ASP
4000
BIO
Effec ve Popula on size
3500
CAM
GCI
3000
GIR
2500
MAL
NIC
2000
ORO
1500
SAA
SAM
1000
SAR
TER
500
VAL
0
0
500
1000
Genera ons Ago
1500
2000
VPS
Selection signatures
Lositan software
Simulation of markers under the neutral model
Detection of outliers having Fstor lower than expected under
a neutral model at that value of heterozygosity
- 456 markers under directional selection
- 629 under balancing selection
What about breed diversity at these loci?
Markers under
directional/balancing selection
0.06
0.02
0.06
0.10
0.25
0.02
0.02 0.06
0.05
Under_sel
Neutral = 50051 markers
Under_Sel = 456 markers
Bal_Sel
= 629 markers
Neutral 1 = 456 markers
Neutral 2 = 456 markers
0.06
Neutral1
0.08
0.02
Neutral
0.010
Bal_sel
0.018
0.02
Neutral2
0.05
0.20
0.35
0.02
0.06
0.010
0.016
Ntr
Pearson’ correlation
between genetic distances
between breeds
Dir
Bal
Ntr 1
Ntr 2
Ntr
Dir
Bal
Ntr1
Ntr2
1
0.85
0.89
0.98
0.99
1
0.70
0.83
0.88
1
0.85
0.88
1
0.97
1
Directional selection
8
●
6
● ●
●
4
●
2
●● ●
●●
●
●
●
●
●
●
● ●
●
●
TER
0
VarExplained(%)
10
12
d=5
PCA − PC1 (%VarExp: 13.509) and PC2 (%VarExp: 4.794)
PCs
●
●
●
●
●●
●●●
●● ● ● ●
● ●●
●
●
●●
●●
●●● ●
●●
●
GIR
●●
●
●
● ●
●
●
●●● ●
● ●●
● ●● ●
●● ● ●
●
●
●
●
●
●●
● ●● ● ●
●●●●
●●
●
● ●● ●● ●●● ●●
●
●
●●●● ● ● ●
●
●
●
●
●
●
●
●
●● ●●● ●●●●
● ● ●
●
● ●● ● ● ●
●●
● ●●
●
● ● ●●
●
●
● ● ● ● ●●
●
●● ●●●●
● ●●●●●
● ● ●●
●
●
●
●● ●
● ● ●●
●
●● ●●● ●●●
●●
●
●
●
●● ●
ORO
VAL
●●
● ●●
●
●●
●●
●
●
VPS
BIO
SAA
CAM
●
●
●●●
● ●
GCI
●●
●
●
●
●
●
●●
●
ASP
ARG
NIC
●
●
●
●
●
●
● ●●●●
● ● ● ●
●
● ●● ●●
●
●
● ●●●●●● ●● ●
●
●●●●
●
●●
● ●● ●●● ●
●● ●●
●● ●
●●
●
●
●
●
●
●●
● ●
●●
SAR
●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
MAL
●●
●
●
●
● ●
SAM ●
●
● ●●
● ●
●
●
●
●
●
Directional selection
0.68
0.62
0.64
0.66
CV error
0.70
0.72
CV error
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
K
SAR
SAM
MAL
GIR
ARG
ASP
NIC
TER
GCI
VAL
BIO
CAM
SAA
ORO
VPS
0.0
0.4
0.8
Directional
Selection
Ancestry
ADMIXTURE Trial K=11
Mostly
Neutral
Alps
Center
South & Islands
Balancing selection
d=2
PCA − PC1 (%VarExp: 0.884) and PC2 (%VarExp: 0.846)
0.4
●
●
0.2
●
●
0.0
VarExplained(%)
0.6
0.8
●
●
●
●
PCs
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
● ●●
●●
●
●●●
●
●●
●
● ●●●
●
●●
●
●●
●
●●
●●
●
●
●● ●
●● ● ● ●● ●
●
●
●
●
●
●
● ●
●
● ●
●●●
●
●
● ●● ●●
● ●●●
●●
●●● ●
● ●
● ●
● ●
●
●●●
●
● ●●● ● ● ●●●●
●●
● ●● ● ● ●
●●
●
●
●
●
●
●
● ●●● ●●
●● ●●
●●●
●●
●●
●● ●
●●●
●
● ●●●
●●●●● ●●●● ● ●
●
●
●● ●
● ● ● ●●● ●● ● ●
●
●
● ● ●● ●
●
●
●
● ●●● ● ● ● ●●● ●●
●
●●●●
●● ●● ● ●
●
●
●
●● ● ●
● ●●
● ●●
●
●
● ● ● ●
● ●●
●● ●● ●●
●
●● ● ●
●
●
● ●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
● ●
●●
●
●
● ● ●
● ●
●
●
●
●
●
●
●
●
●●
● ●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
CAM
NIC
GCI
ARG
SAR
ASP
MAL
VAL SAA
VPS
GIR
BIO
SAM
●
●
●
TER
●
●
ORO
●
●
●
●
●
●
●
●
●
●
●
Balancing selection
0.65
0.70
CV error
0.75
0.80
CV error
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
K
SAR
SAM
MAL
GIR
ARG
ASP
NIC
TER
GCI
VAL
BIO
CAM
SAA
ORO
VPS
0.0
0.4
0.8
Balancing
Selection
Ancestry
ADMIXTURE Trial K=11
Mostly
Neutral
Alps
Center
South & Islands
0.38
0.410.34
0.38
Correlation between
heterozygosities of breeds
0.34
0.38
0.34
0.38
0.42
0.38
Neutral2
Neutral2
Neutral2
Neutral
Neutral
Neutral
Neutral
Neutral1
0.42 0.39
0.36 0.39
0.34
Neutral1 Under_sel
Under_sel
0.34
0.26
Neutral1
0.26
0.38
0.36 0.39
0.36 0.39
0.35
Neutral1
0.34
0.38
0.35
0.41
0.36
0.39
0.36
0.39
0.26
0.380.26
0.32
0.34
0.39
0.32
0.38
Neutral
0.39
0.41
0.36
0.39
0.36
0.39
1
0.26
Directional
Balancing
Neutral 1
0.32
0.26
Under_sel
Neutral Under_sel
Directional
0.380.26
0.87
0.32
0.38
1
0.26
0.39
0.39
Bal_sel
Neutral2
0.42 0.39
0.41 0.35
0.39
0.39
0.38
0.34
0.410.34
0.34
0.38
Bal_sel
0.35
_sel
0.41
0.41 0.35
0.36 0.39
_sel
0.41
0.38
0.42
0.35
Balancing
Neutral 1
0.33
0.99
0.14
0.85
1
0.33
1
Correlation between
heterozygosities of individuals
0.35
0.40
0.40
0.50
0.30
0.30 0.35 0.40
0.30
Bal_sel
Under_sel
0.30
0.40
0.50
0.15
0.25
Neutral
Directional
Balancing
0.35
0.15 0.25 0.35 0.45
Neutral
0.45
Neutral
Directional
Balancing
1
0.76
0.66
1
0.33
1
IMMAGINE
Conclusions
IMMAGINE
SNPs vs others
• Much higher level of resolution (many thousand vs a
few markers)
• Robust and non homoplasic
• Easier comparability across projects and data
merging
• Suited to genome wide analyses (ROH, Ne, Selection
signatures, GWAS, breeding applications)
• However panels should be carefully prepared and
evaluated (ascertainment bias)
Italian goats
• Little or no inbreeding.
• Variable level of admixture.
• Some distinct breeds: Girgentana, Teramana, Orobica,
Maltese.
• Low Ne nowadays (bottlenecks, breeding management),
higher in the past.
• Geographic partition of diversity at small geographic scale
(North-South and East-West in the Alps).
• Markers under selection are valuable for conservation
decisions
• Neutral marker diversity is a reasonably good proxy of
diversity of markers under directional selection
Breeding
• Breeding will be more and more guided by
molecular analyses if cost continues to decrease
• Methods customised to populations (small vs
large pop. improvement, inbreeding control,
maintenance of diversity)
• Knowledge of population structure is needed for
any kind of application to avoid false positives
30
International Networking
ADAPTMAP
Goat Adaptmap
Traditional and novel approaches to
study adaptation genomics:
Selection signatures
Spatial analysis
Enriched SNP panel
Detection of new variation
To chacterize the study
Population studies based on Mutation
Classification
Mutation effect on protein structure and
function used as a tag for adaptation
Alessandra Stella
NextGen project
Pierre Taberlet
Francois Pompanon
Next generation methods to preserve farm animal biodiversity
by optimizing present and33 future breeding options.
Interdisciplinarity and
training
GIS
Training
34
Future challenges
Copy Number Variations (CNVs)
Mefford and Eichler 2009
See following presentation of Fernando Garcia
Final consideration
• Very fast molecular tool development.
• Faster than our capacity to understand.
Under these circumstances any loss of
diversity before characterization is a loss of
unvaluable opportunity for science and
agriculture
36
ACKNOWLEDGMENTS
Marco Milanesi
Elia Vajana
Lorenzo Bomba
Licia Colli
Goat farmers
International Goat Consortium (SNPChip)
ASSONAPA (Italian small ruminant breeder
association)