1. Participatory crop
improvement for local /
regional adaptation and
produces
Isabelle Goldringer,
INRA
Rennes, Diversifood Final Congress, December
10-12, 2018

2. • Substituting chemical inputs by biological regulations induces:
- more diverse farming practices
- more contrasting environmental conditions
= Genotype x Management
x Soil x Climate
=> decentralized breeding
Issues in breeding for sustainable farming
and local adaptation and produces (1)
Genotype x farming system interaction
(Murphy et al. 2007)

3. • Climate change and less protective conditions lead to:
- more variable and unpredictable climate and environments
= Need for buffering capacity
Within field crop genetic diversity = lever for productivity and
resilience
=> breeding for populations
Issues in breeding for sustainable farming
and local adaptation and produces (2)

4. • Varieties must be adapted to:
- farming practices, local conditions
- the processes, produces, market chains
= Incorpore the knowledge of the actors,
= utilize a contextualized, inclusive innovation process
=> participatory approach
Issues in breeding for sustainable farming
and local adaptation and produces (3)

5. 5
Participatory Plant Breeding
• Participatory Plant Breeding involves scientists, farmers, and others
(consumers, processors, extensionists,…) in plant breeding research.
• Participatory approaches: the processes used to develop new
technologies can achieve a range of material and socially
transformative outcomes.
≠ transfer-of-technology model that seeks efficient processes
to generate standardized, uniform outcomes (Jones et al 2014).
=> New methods needed:
* based on participatory approaches
* adapted to decentralized on farm trials
to manage, breed and adapt crop genetic diversity, embedded in
local/regional food systems

6. 6
DIVERSIFOOD objectives for PPB (WP3)
• Establish new populations & evaluate the
relative interest of the different methods used
• Develop or adapt methods and user-friendly
tools
• specifically designed for on farm decentralized
participatory breeding
• Apply a panoply of breeding strategies and
compare them to identify the most promising

7. 7
New populations created for many crops
species
- A large number of new populations developed for many
crops: tomato, broccoli, carrots, onion, bread and durum
wheat, einkorn, barley, maize, white lupin, faba bean and
buckwheat

8. A range of methods to create relevant
diversity for PPB
• Landraces and old varieties (ex: ITAB, RSP):
– Large diversity available, adaptive
potential to low input and agroecological
conditions, organoleptic and nutritional quality
– Needs further adaptation
• Mixtures of landraces and old varieties (ex: INRA, ITAB,
RSR):
– Allows to combine interesting traits from different varieties, easy
and rapid
– Limited recombination in selfing mixtures, risk of competition, needs
management
• Bi-parental crosses (ex: INRA, PSR):
– Creation of new genotypes combining advantageous traits from
both parents
– Crossing plants of self-pollinated species is feasible on-farm by farmers,
but is time consuming and requires dedicated skills.
8

9. A range of methods to create relevant
diversity for PPB
• Landraces and old varieties (ex: ITAB, RSP):
– Large diversity available, adaptive
potential to low input and agroecological
conditions, organoleptic and nutritional quality
– Needs further adaptation
• Mixtures of landraces and old varieties (ex: INRA, ITAB,
RSR):
– Allows to combine interesting traits from different varieties, easy
and rapid
– Limited recombination in selfing mixtures, risk of competition, needs
management
• Bi-parental crosses (ex: INRA, PSR):
– Creation of new genotypes combining advantageous traits from
both parents
– Crossing plants of self-pollinated species is feasible on-farm by farmers,
but is time consuming and requires dedicated skills.
9

10. A range of methods to create relevant
diversity for PPB
• Landraces and old varieties (ex: ITAB, RSP):
– Large diversity available, adaptive
potential to low input and agroecological
conditions, organoleptic and nutritional quality
– Needs further adaptation
• Mixtures of landraces and old varieties (ex: INRA, ITAB,
RSR):
– Allows to combine interesting traits from different varieties, easy
and rapid
– Limited recombination in selfing mixtures, risk of competition, needs
management
• Bi-parental crosses (ex: INRA, PSR):
– Creation of new genotypes combining advantageous traits from
both parents
– Crossing plants of self-pollinated species is feasible on-farm by farmers,
but is time consuming and requires dedicated skills.
10

11. A range of methods to create relevant
diversity for PPB
• Composite Cross Populations (CCPs)
and mixtures of crosses (ex: INRA, FiBL,
ORC, RSR, LBI)
– High diversity and recombination, large
potential for adaptation and mass
selection => broad genetic basis population
– Heavy to carry out and difficult to design
• Open pollinated populations (ex: ITQB & IPC):
– In out-crossing species, allows to combine several parents
in one or a few steps
– Difficult to apply mass selection on both parents, requires
optimization of the within-population diversity
11

12. A range of methods to create relevant
diversity for PPB
• Composite Cross Populations (CCPs)
and mixtures of crosses (ex: INRA, FiBL,
ORC, RSR, LBI)
– High diversity and recombination, large
potential for adaptation and mass
selection => broad genetic basis population
– Heavy to carry out and difficult to design
• Open pollinated populations (ex: ITQB & IPC):
– In out-crossing species, allows to combine several parents
in one or a few steps
– Difficult to apply mass selection on both parents, requires
optimization of the within-population diversity
12

13. 13
Definitions and description of PPB
varieties and populations
Deliverable D3.3 & Innovation Factsheet #2

14. Development of CCPs of white lupin at
FiBL
• 2 strategies for CCP creation:
– With or without selection for Colletotrichum
resistance
14

15. CCPs, dynamic populations and
evolutionary breeding
• Comparison of a CCP and a dynamic
population (mixture) of bread wheat
at INRA & ITAB:
– See Estelle Serpolay’s poster
• Test of selection methods in wheat CCPs
at LBI:
– Rust, seed size, plant height, deep sowing vs
normal sowing
• Evolutionary breeding
with a bread wheat CCP at RSR:
– See Bettina Bussi & Matteo Petitti presentation
15

16. Multi-trait approaches to PPB in maize
• Combining agronomic, quality and genetic
information to optimize crosses at ITQB-NOVA & IPC
– Crosses chosen to increase tocopherol and total
phenolic content , minimize viscosity
while maintaining genetic
diversity
• Decision tools for improving underused Portuguese
maize landraces at ITQB-NOVA & IPC:
– see ML Alves poster
16

17. 17
• On-going PPB on bread wheat since 2006 in France (INRA
& RSP). First 10 PPB population-varieties evaluated on
farm for agronomic performance, nutritional and
organoleptic qualities.
In the middle: Renan, short,
poor weed competitor, low
bread making quality
Aside and below: PPB
varieties, tall, robust, good
weed competitor, lodging
resistant, good bread making
quality
Original and interesting populations-varieties
developed through PPB at INRA & RSP
PPB varieties showed
higher level in Mg, Zn,
Ca, Cu and K than
commercial varieties.

18. Instability
(VC %)
Farm Year +
Year x Farm
R Farm Year +
Year x Farm
R
PPB
Populations
1% 24% 6% 23% 28% 17%
Commercial
varieties
0% 35% 5% 37% 35% 26%
• First 10 PPB population-varieties evaluated on 6 farms x
2 yrs for agronomic performance
Wheat PPB populations are more stable
Protein content Yield
Thèse G van Frank (Dec 2018) + see Gaelle’s poster
a 32 qx/ha
29 qx/ha

19. Collective assessment in PPB
• Participatory tomato breeding at Arche
Noah:
– Networking and cooperation between different
stakeholders (farmers, consumers, retailers,
chefs, scientists) => to define breeding
strategies.
– Collective breeding activities: participatory
assessment of quality traits.
19

20. Row
1
Row
3
Row
4
Row
5
Row
6
Row
2
Col 1 Col 2 Col 3 Col 4 Col 5 Col 6
Innovative statistical methods, designs and
tools
20
• Participatory on-farm experiments means:
– Numerous but small trials (few plots / trial)
– Few replication within and among farms
– Farmers evaluating different populations
– Broad range of traits selected
⇒ Innovative methods for statistical analyses and data mining
specifically suited to PPB (RSP, RSR, IPC, ITQB, INRA)
• M1. Non parametric multivariate approaches
for agronomic and nutritional traits
• M2-M3. Multivariate and clustering approaches
for molecular and quality data
• M4-M5. Spatial analyses with optimized randomization & incomplete block
designs for agro./nutri. traits
• M6. GxE analyses for agro./nutri. traits
• M7a-b. Models within the Bayesian framework
for agro./nutri. traits
• M8. Analysis of network of seed circulation
• M9. Multiple factors analysis for sensorial data

21. 21
Technical booklet #3 – “Methods and tools for decentralized
on farm breeding”
⇒ A decision tree including all possible analyses organized
according to the type of trait, their objectives and
the experimental capacities (RSP, INRA, RSR, IPC, ITQB)
Innovative statistical methods, designs and
tools

22. PPBstats : An R package for PPB
statistical analyses
• Objectives:
– (i) to have a single package capable of performing all
analyses required for PPB programmes with
comprehensive documentation
– (ii) to create a community working on PPB programmes
in order to improve the package, exchange on how to
process data from PPB programmes and develop good
practices.
• Methods:
– Freely available package based on the R software
– Collaborative plateforme on github
22

23. PPBstats : An R package for PPB
statistical analyses
• Worlflow of analyses on data collected during PPB
programs at four levels: network of seeds management,
agronomic and nutritional trials, organoleptic tests and
molecular analyses.
• A website dedicated to PPBstats with full documentation
and a comprehensive tutorial to collaborate and use the
package can be found at:
https://priviere.github.io/PPBstats_web_site
• Use and contribute !!!
23

24. Key features and outcomes of PPB
• New populations developed and adopted by the farmers
• Complementary to ex-situ conservation: create and renew crop
genetic diversity, increase use of landraces
• Relies on a collective organisation
– Very inclusive multi-actor approach with various types of
stakeholders involved: researchers, farmers, breeders, processors,
cookers / chefs and consumers
– Generic tools and methods for supporting collective selection /
management of diversity
• Building capacity and empowering farmers to enact further
change

25. Conclusion
=> Increase system resilience :
- For each farmer (resilience of population-varieties)
- At the collective level: seed control and decision making less
concentrated and genetic variability is distributed
=> A new paradigm for research: collaboration between
research and citizens
– Elaborate original and complex research questions
where inter-disciplinary basic and applied research
are interlocked
– Give more sense to research
– Stimulate citizens’ interest and involvement into
research and identification of research needs
V Chable, INRA Rennes
On Raphael Balthassat’s farm in
2015: his PPB variety Savoysone

26. Bottlenecks and levers for a
transition
• A long term process.
• Scaling up and/or making these small scale systems sustainable
is a matter of training / organizing / funding.
Need to:
– Train farmers’ organisation facilitators: they are key to the collective
organization
– Increase sharing of tools and methods for PPB among local groups
through common framework and perenial platforms
– Develop multi-stakeholders participatory research platforms to allow
for recurrent funding
– Find sources of funding to make local community seed banks
sustainable
– Create a status for varieties derived from PPB

27. 27
Thanks to the DIVERSIFOOD partners
(WP3)
Thank you for your attention !
PPB bread wheat population-variety Japhabelle

28. Ten bread wheat varieties derived from PPB
Nom Origine Type
Saint-Priest F Mercier, Maine-et-Loire Issue d’une variété Suédoise (Progress)
Rouge du Roc JF Berthellot, Lot-et-Garonne Sélection massale dans Rouge de Bordeaux
Pop dynamique 2 F Mercier, Maine-et-Loire Mélange de 3 variétés de pays et 2 variétés + récentes
Mélange-5
Bourguignon
B Ronot, Côte d’Or Mélange de 11 variétés de pays locales
Mélange du Sud-Ouest JF Berthellot, Lot-et-Garonne Mélange de ~ 20 variétés de pays locales
Savoysone R Balthassat, Haute-Savoie Issue d’un croisement entre 2 variétés de pays
Rocaloex R Balthassat, Haute-Savoie Mélange de 11 populations issues de croisements
Mélange-1 13 pops B Ronot, Côte d’Or Mélange de 13 populations issues de croisement
Dauphibois C Dalmasso, Isère Mélange de ~25 variétés de pays, populations issues de
croisements et variétés récentes
Japhabelle JF Berthellot, Lot-et-Garonne Mélange de ~25 populations issues de croisements et
sélectionnées