1
Breeding without breeding with only forests and
combined seed orchards/breeding populations 

• Dag Lindgren and Xiao-Ru Wang
• Swedish University of Agricultural Sciences
• and
• Umeå University
• Seed orchard Meeting
• IUFRO WP 02-09-06
• Korea 090909

http//www-genfys.slu.se/staff/dagl/korea09/
2
Acknowledgements

• Yousry El-Kassaby
• Anders Fries
• Takeshi Torimaru
• Matti Haapanen

3
A system with only clonal seed orchards and
forests
Forests
Selection
Seed orchard
4
Actually this system is used for half of Swedish
plantations today
Forests
Seed orchard
5

• Clonal seed orchards (grafts) and forests are
  managed and paid for anyway by forestry (sunk
  cost, not a cost for breeding).
• Recombination is done anyway in seed orchards and
  the plustree hybrids are planted anyway in
  forests.
• Phenotypic selection can be done fast and
  efficiently without any measurements in young
  forests by subjectively choosing good trees and
  without any knowledge of the ancestry of the
  trees.
• Phenotypic selection in plantations has the
  potential to work well even in advanced
  generations.
• No permanent organisation for managing clonal
  archives, controlled crosses, field tests and
  measurments is needed (BwB Breeding without
  Breeders). The investment in genetic evaluations
  to make a new seed orchard need to be done only a
  few years before the establishment of the new
  seed orchard instead of supporting long term
  breeding 35 years before material for new seed
  orchards are obtained as done in Sweden today.

6
But what about relatedness?
7
Mass selection (phenotypic selection) offer a
good and near optimal trade off between gain and
gene diversity
Gain
0
1
Modified From Lindgren and Wei (1993) and
developed by Lindgren and Wei 2006
8
Phenotypic selection in a progeny plantation
gives a near optimal gain at the relatedness
level obtained (diversity)
9
Phenotypic selection (mass-selection) in improved
forests has problems, e.g.

• Contamination with unimproved trees and pollen
  reduces gain.
• Relatedness will not be controlled or known and
  perhaps large.

These problems can be overcome if trees are
phenotypically pre-selected in the forest and the
ancestry of the pre-selected trees clarified by
markers. A variant of Breeding without Breeding
as coined by Yousry El-Kassaby.
10
Simple Breeding without Breeding Cycle
Forest
Select considering
Seed orchard
11
Do markers work?

• We made a study where the trees in the forest was
  seeds in a seed orchard harvest (the same as
  presented by A Fries at this meeting).
• The mothers (seed orchard clones) were known, but
  pretended unknown.
• The study Wang X-R, Torimaru T, Lindgren D and
  Fries A 2009. Marker-based parentage analysis
  facilitates low input breeding without breeding
  strategies for forest trees. TGG in press.

12
Material methods

• Seed orchard with 28 clones (Västerhus)
• 8-12 seeds/clone, 305 seeds in total
• Nine selected SSR (microsatellite) markers
• Parent-pair analysis with both parents unknown
• Simple exclusion analysis assuming null allele

13
The simple system with only seed orchard and
forest. This is Västerhus seed orchard the
pollination year of seeds studied.
Västerhus spring 2006. Photo Yoshinari Morogushi
14
Commercial plantation with seeds from the orchard
Photo Dag Lindgren
15
Selections are done in the forest from the seed
orchard. Needles are collected for ancestry of
pre-selected trees
Photo Takeshi Torimaru 0806
16
Results

• For the offspring from the seed orchard, clones
  both parents can be determined for 93-97
  offspring.
• Thus, in principle BwB is technically possibile.
• The cost for the marker analyses is reasonable
  (given the competence exists).

17
Old trial with crosses from some of the Västerhus
clone and others

• A pilotstudy with 35 trees was done in a progeny
  test where some progenies of Västerhus clones
  were planted.
• Few needles lt9 m! That made collection
  problematic. It is practical to select trees at
  an age when fresh needles lt5m!
• All trees analysed, which did not origin from
  Västerhus clones, could correctly be excluded.
• Ancestry for a few trees did not agree with
  documentation. It would be compatible with 10
  erroneus pedigree in progeny-trials. We found it
  not reliable enough to go on with.
• But it illuminates that pedigreed tests may not
  be quite certain, thus the advantage over BwB is
  less than calculated.

18
Experiences and development

• The required competence for marker analyses and
  statistical analyses is high. We do not recommend
  it without involvement of an experienced
  molecular lab and some skill in the
  mathematical/statistical analyses.
• Theoretically there exits a large number of
  marker loci (SSR), but in practice brand new
  suitable loci (little trouble with null-alleles
  and high polymorphy) are not easily identified.
• There is a technical development and the methods
  can be improved, but we do not recommend to take
  for granted that this BwB application functions
  well for seed orchards with more than 50 clones
  even in a decade.
• The pollen contamination in Swedish Scots pine is
  near 50, this BwB application would function
  better when pollen contamination is lower.

19
Comments about this simple BwB

• If breeders start using BwB, the step to other
  uses of the same samples of DNA is short. BwB may
  open the door to other applications of molecular
  breeding.
• Simple BwB produces trees with uncertain breeding
  values (phenotypic selection). There is a desire
  to have tested clones in advanced seed orchards,
  but that is probably not optimal, so do not be
  afraid of using untested clones in sufficient
  numbers.
• BwB means a wait for good pollen production,
  artificial crosses can be made earlier.

20
Progeny test

• It is possible to link good trees in the forest
  to parents in a seed orchard. Thus it is also
  possible to determine breeding value of the
  orchard clones for ability to give good
  offspring. Such a progeny-test seems to cost
  only a fraction of conventional progeny-testing
  allthough less accurate.

21
Swedish long term breeding

• Sweden has since six decades invested much in
  field trials, breeders and breeding organisation.
  Thanks to that, the new seed orchards have three
  time more genetic gain than the first batch.
• Swedish long term breeding is mainly based on
  within family selection with all parents equally
  represented, thus parental breeding values does
  not matter. This simple BwB system with
  phenotypic pre-selection, followed by equalising
  parental contributions is similar to the Swedish
  system.

22
Comparison

• We compared this simple low-input system with
  current Swedish long term breeding. Under our
  assumptions the low input system produced
  somewhat less than half the annual gain, but
  annual costs can be cut almost 90!
• Sweden is (hopefully) willing to pay the higher
  cost for the more powerful and better controlled
  high intensity breeding for valuable crops (Scots
  pine and Norway spruce).

23
BWB in minor species?

• Low need for seeds justifies low-input solutions.
• The pollen contamination is usually low or
  absent, which makes BWB much more effective.
• But I guess that grafted seed orchards are
  complicated for minor species, it is preferable
  with seedling seed orchards (gene resource
  populations combining different objectives).
• BWB in its present form can only handle clonal
  seed orchards. But in the future we may identify
  grandparents with markers, and when BWB can be
  used for seedling seed orchards.

24
BwB with grandparents

• It seems considerable more effective to keep
  grandparents under control than parents. Thus it
  seems desirable to develop BwB systems where
  pedigree is tracked to grandparents.
• Lindgren D, Danusevicius D Rosvall O 2008.
  Balanced forest tree improvement can be enhanced
  by selecting among many parents but keeping
  balance among grandparents. Canadian Journal of
  Forest Research 38(11) 27972803.

25

• It can be argued that Scots pine in southern
  Sweden is a candidate for converting the current
  high input breeding to low input BwB. The recent
  planting is low mainly because of damage by
  animals (moose) and the plant number supported
  hardly justifies the current breeding effort.
• But forestry wants seed orchards in reserve
  because demand may increase.
• Could these seed orchards be simple BwB?

26

• Some consideration starting BwB with F1-hybrids
  with plustree parents in an advanced system like
  Scots pine in southern Sweden
• To make BwB powerful in a system whith F1-clones,
  many untested phenotypic F1-clones should be in
  the seed orchard. The final selections from the
  forest should not include offspring from many
  F1-clones to get gain from among family
  selection.
• The F1-clones will be related, but it is not
  desirable with related clones in a seed orchard
• Simple BwB can only handle 50 clones.
• Advanced seed orchards should have rather few
  clones for high gain.
• If the aims - 1) preserving diversity for
  breeding and 2) high gain immediatly for forestry
  - are served by the same seed orchard, both will
  loose.
• How can those problems be handled?

27
An simple orchard-forest BwB design for advanced
generations

• Have orchards with the major contributions from
  the 20 best clones. But in addition have say 100
  clones with very few (two) copies. The production
  loss will be very low and partly balanced by more
  diversity and flexibility.
• The extra 100 clones can be spatially close.
• Selective harvests are done from these extra
  clones in batches of 40 and special forests (not
  large) established and documented.
• For next seed orchard BwB selection (phenotypic
  pre-selection followed by marker analyses and
  final selection) is done in the special gene
  resource forests besides the ordinary forests. In
  the gene resource forest the rare clones will get
  good representation.

28
A system with only clonal seed orchards and
forests,but more clones
Forests
Seed orchard
Selection
29
Flexibility

• The suggested seed orchards can also be a
  starting point for OP-progeny tests with
  subsequent BwB selections.
• The seed orchards can be established today, and a
  decision can be made in 15 years, when they are
  adult, if it will be OP-progeny-tests or simple
  BwB (commercial forests planned for selection).

30

• Thanks for attention