Low Input Tree Breeding Strategies

1
Low Input Tree Breeding Strategies

• Dag Lindgren
• Department of Forest Genetics and Plant
  Physiology, Swedish University of Agricultural
  Sciences, Umeå
• Sweden
• September 2, 2002, China

2
Scientists focus high input
Because of training and ambitions the attention
is mostly here
3
High-input techniques

• Genotype testing,
• Controlled crosses,
• Known pedigrees,
• Orchards intensively managed exclusively for
  seed production,
• Grafts for seed production.

4
High input

• Scientific rewards and fancy journal papers,
• Collegues admiration,
• Non-sloppy well organised programs,
• The main driving force of human phsyc,
• Focus for teaching in schools and textbooks at
  Universities.

5
High input and low input
Attention should be here for low-input strategy
6
Low input situations

• Poor
• Unstable organisation
• Local control
• No specialists
• Minor program
• Lower tiers of breeding population (main, but not
  elite)

7
Low-input techniques

• Selection on phenotypes instead of testing of
  genotypes,
• No records of tree ID or pedigree,
• Wind pollination,
• Seed production in stands used for other
  purposes,
• Cheap plantations created for future seed
  production and long term improvement.

8
Low-input techniques

• Thin stands rather intense to get better pollen
  and take seeds from best trees

• ButDepend on predictions of inbreeding,
  coancestry and diversity replacing pedigree.
  Predictions may fail. and are generally not even
  made yet.
• Note that there is no strict limit between high
  input and low input techniques!

9
Phenotypic selection

• No tree identities required,
• No computer required,
• No strict objective measures required,
• Transparent (no black box)
• Can be executed immediately in field,
• A type of selection forwards,
• Can also be called mass-selection,
• Similar to how Nature selects, thus sustainable
  and environmental friendly.

10
Phenotypic selection

• Depending on predictions for control of
  accumulation of relatedness,
• Can be diversity use efficient,

11
Maximising gain at a given diversity by selection
in infinite normal distributions. h20.25 and
P0.1
Combined indexestimated BV (maximizes gain)
Phenotypic selection (easy)
Gain
Between family(exhausts diversity)
Within family(preserves diversity)
Diversity
Note that phenotypic selection is on the
optimising curve, thus no other selection results
in higher gain without sacrifying diversity!
Modified From Lindgren and Wei 1993
12
Testing is doubtful for low-input breeding

• more complicated,
• more demanding on temporal and organisational
  stability,
• a considerable long time investment requires
  trust on that the results would be used,
• selection forward (untested) is often found to
  offer as high gain as selection backwards (means
  progeny-test doubtful effectiveness),
• Progeny-testing often not competitive.

13

Restricted selection for Phenotypic and Breeding
value (combined index, conciders both individual
and family) in a population created by 20 parents
with family size 20, h20.5. Points correspond to
restriction intensity. Simulation (POPSIM).
14

30
Phenotypic
Unrestrictedbreeding valueestimate
24
Gain
Balancedselection
18
12
4
6
8
10
12
14
Effective number (Ns)
15
Note

• Phenotypic selection as good as restricted
  selection for breeding value compared at same
  gene diversity!
• Unrestricted breeding value selection gives a
  higher gain, but at the cost of a lower gene
  diversity!
• Remember Breeding value is based on a combined
  index of an individual and its sibs used for
  selection forwards. It does not refer to parental
  ranking.

16
Restricted selection for Phenotypic and Breeding
value for several generations One and five
generations of selection in a population with a
family structure, h20.5, family size 20
Breeding value
Phenotype
110
100
5 generations
90
80
70
Gain
60
50
40
30
20
1 generation
10
0
0.1
0.2
0.3
0.4
0.5
Loss of gene diversity
17
Note

• Phenotypic selection is compatible also in a
  multigeneration program
• For unrestricted selection for breeding value the
  genetic variation get exhausted after a number of
  generations, and in the long run the gain than
  with phenotypic selection is higher
• However, if breeding population large and
  heritability not very large, this exhaustion
  takes long time.

18
One and five generations of restricted selection
in a population with a family structure, h20.05,
family size 500. Low heritability and large
families favor combined index
Phenotype
Breeding value
40
5 generations
30
Gain
20
10
1 generation
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Loss of gene diversity
19
Development of Gain and gene diversity over five
generations of selection in a population with a
family structure, h20.05, family size 500 for
three selection strategies.
20
Selection strategies
21
Fertility variation matters for accumulation of
relatedness over generations
The development over generations in a closed
population of 154 teak trees based on their
observed fertility variations (Bila et al. 1999)
22
Grafted seed orchards not low-intensity

• large investment,
• long-term investment,
• trust in use of product,
• large scale (management, pollen),
• special technical competence required,
• particular land use,
• organisational stability over time required,
• selected fathers, but....

23
Gene resource plantation

• Looks and is managed similar to a "normal"
  plantation,
• Limited need of specialised competence and
  organisational stability,
• Multiple use (options for seeds and improvement,
  wood, conservation...),
• Can function as seed collection area,
• As cheap trees may even be cut for seed
  collection,
• Can be close to local organisation, enterprise
  and people,

24
Gene resource plantation

• Robust to disasters or neglectance,
• Small investment, thus limited loss if interest
  lost
• Optional plant identity, but,
• Seeds for commercial use and replacing can be
  improved by thinning and by harvesting the best
  trees,
• Renewed by wind-pollinated seeds mainly from the
  plantation,
• Coancestry controlled by predictions, sufficient
  numbers and seed parent contribution control.

25
Will "we" be needed?

• Need to optimise "low-intensity" strategy and
  tactics (almost not done yet)
• Less control (wind-pollination, unidentified
  trees etc.) means much more thought and
  predictions needed
• The reduced need of competence is "technical
  competence and field competence. Not competence
  of breeding scientists
• Low-input should replace No-input, not High-input