Increasing Genetic Gains for Food Security in the Developing World: Identifying innovative solutions to legumes, roots, tubers and bananas breeding

Crops are Cotton, soybean, pea, lupin, broccoli, apple. Most important traits are resistance against diseases and pests, nutient use efficiency, abiotic stress tolerance, product technological quality, protein quality,  

Constrains: unsecure weather events (drought, flooding, flooding, frost) and massive attack of pest and diseases

I work mainly with cowpeas, sweet potatoes and cabbages.

Groundnut (Peanut)

Groundnut (Peanut)

Drought tolerance/resistance,  diseases (particularly leaf spots, rosette & aflatoxin) reistances and oil content are focus traits

Phenotyping protocols and facilities for drought and aflatoxin are not to the level where acurate predictions can be made. Screening are usually made under natural conditions as facilities for phenotyping under controlled conditions are limited or lacking. Most of the breeding programs do not have the necessary facilities for routinely analysing oil content and quality in large number of breeding lines.

I'm working with peanut and the three most important phenotypes we're looking for in our breeding program are drought tolerance, oil quality and pre-harvest aflatoxin contamination resistance. For drought tolerance, the peanut research community has been extensively screening large panel of genotypes for identifying surrogates that breeders could use when selecting for drought tolerance but I think it's still challenging to make decision when it comes to choose surrogates for drought tolerance in peanut. Indeed, correlation entre surrogates and yield and/or yield components may depend upon drought stress severity and even with standardized phenotyping protocols the magnitude of the correlation was not always very similar.

For aflatoxin contamination resistance, it's really challenging to phenotype genotypes under uniform pattern of the fungus. Then in most of the studies, there's a large variability between replications for a given genotype. Therefore, it is difficult to select for resistant genotypes due to the nature of the infestation.

Dry bean Phaseolus vulgaris.  Critical traits are end user traits,  seed size, color, shape,  lack of defects, canning quality,  follow this with agronomic traits,  maturity,  lodging resistant,  uniform dry down,  third group is specific disease resistance traits without which the variety will fail either in seed multiplication or production.

Peanut.

Disease resistance, drought tolerance, quality

phenotyping for aflatoxin (a quality issue) is very difficult

Crops: Groundnut/Peanut, Common bean, soybean and rice. The constraints other than yield are the expected ones: biotic and abiotic stresses. There are important considerations for downstream usage (e.g. oil content, size, etc.) that affect their value, but the ability to grow without extensive inputs is important.

Dry beans. Critical phenotypes besides yield are seed quality traits (shape, size, color, canning, etc.), plant architecture, maturity and plant drydown. Then you have all the traits related to both biotic and abiotic stress.

I work with potato, sweetpotato, maize. Storage capacity (potato, sweetpotato); disease resistance; abiotic stress tolerance.

Crop: common beans in Central America; Constraints: diseases (many); production system appropriate architecture; local control of seed industry by small-holder farmers. Key phenotype constraints: accurate genotyping to preclude time-consuming phenotyping.

Groundnut (Peanut):

Traits: Diseases (Groundnut rosette disease and leaf spots); quality (aflatoxin and high oleic) and drought

Peanut. Disease resistance (foliar diseases, TSWV, sclerotium), drought tolerance, aflatoxin. Drought tolerance is hugely complex, still poorly defined in the crop, environment dependent. To tackle drought, it is necessary to have basic understanding of the environment, as well as the anatomical/physiological traits that may help good yield under hydric stress - the lack of this leads to wasted time and resources. Aflatoxin phenotyping is also difficult to reproduce. 

1. We work with potato & sweetpotato. Most important phenotypes: a) Root & Tuber set morphology and adaptation; b) taste often with considerable variation from region to region / subregion to subregion, nutrient density  & processing attributes such as low sugar; c) sweetpotato virus disease resistance; potato late blight resistance, potato leafroll virus (PLRV) resistance, Potato virus Y (PVY) resistance, Nematode resistance.

I work with commoon bean;mostly with dry beans but also with snap beans. The most important phenotypes in my program is genetic resistance to diseases caused by pathogens with extensive virulence diversity. These pathogens include those that cause the rust, anthracnose, angular leaf spot,and halo blight diseases. These r some of the most devatating diseases of common bean in Africa and the Aericas. These virulerence of these pathogens may change from one year or location to another. When new races appear varities with single disease resitance genes are at most risk and often loose their resistance. Yield and pod and seed quaity loses may be very high.The phenotype of these diseasesin fairly well known.he main constraint is a genetic reistance strategy that uses narow gentic base solution (such as single genes) that ignores the diversity of the pathogen and of the host.

peanut (groundnut) and cowpea.  foliar and soilborne fungal diseases, aflatoxin contamination, and seed composition/quality are important traits.  The variability in aflatoxin contamination, even within same genotype and environment, constrains selection for resistance.

Decentralized participatory breeding including farmers and value chain in the target environment!

Easy applical tools for resistance screening.

Improved understanding of plant microbe interaction and their utilization in plant breeding 

Not so much with breeding, but with improving yields; like Monika, the inter-action with soil biology is important, and work on biological pest and disease control, and on resistant/ tolerant varieties.

The progam was able to integrated farmer participatory variety selection approach in the preformance evaluation of advanced breeding lines. This helped in speeding up the release of best bet farmer preferred varieties and adoption. Besides, the program has moved from paper based to digital breeding process managment (trial design, data collection, analysis and storage) for enhanced efficiency

Tools other than directly selecting for the desired phenotype would be molecular markers linked to the trait. In beans we have been using markers now since the early 1990s to select for major disease resistance genes. Markers for quality and agronomic traits have not been as useful as agronomic traits are easy to score and require seasonal field selection - and variability in quaility traits over locations/ years have prevented their adoption.  New tools to indirectly measure quality traits such as NIR visual measures appear promising.

We have integrated molecular markers into breeding programs. Technological advances that reduced cost per data point, and better understanding of how to correct for tetraploid genetics was fundamental. Sequencing of the diploid ancestors of peanut has taken things to a different level.

We have integrated wild species derived allotetraploids to breeding programs. Good partnerships were crucial, and realistic expectations of the long time frame to get to improved peanut lines.

Also the routine use of lab-based detached leaf assays for disease resistance (for rust and late leaf spot), as a complement to field evaluations. This has improved the accuracy of phenotyping for these diseases.

Funding from the peanut industry has been fundamental for sucesses so far.

Critical tools are still selective phenotyping both in the field and greenhouse, but more recently the aid of specific molecular markers have been useful, but with limitations. In dry beans we have many succesful examples of use of molecular markers linked to disease resistance genes. However, most of these are related to monogenic traits (Mendelian/qualitative inheritance). There are very few succesful examples for multigenic/complex traits, which is where the real help is more needed. Even with better markers, genomic sequences, and many other gains in molecular biology, I still don't see this well addressed/applied in breeding programs.

Use of genomics is critical to accelerate breeding, understanding the diversity of the germplasm, and revealing key loci and alleles associated with domestication

Guatemala (beans): detailed genotyping of germplasm base to determine variation in the program that can assist parent selection; markers loosely linked to disease resistance genes, but a group of the markers are only genepool/market class specific.

Classical breeding and farmer participatory breeding

In our research program (using wild Arachis to introgress resistances into the peanut crop),  partnership with various breeding programs has been vital - for validation and utilization of resources. The use of genotyping platforms has sped up marker usage and MAS.

(i) Decentralizing breeding 5-6 relative small units comprising 1 scientist 1 technician and operating funds to cross and select consider local farmer / gender needs under an umbrella of a hub which fosters: population improvement (large scale recombination), standardized data management & analysis, testing & validation of innovations in applied breeding populations as well as meetings across units, knowledge exchange, & seed exchange across breeding units, (ii) accelerated breeding schemes replacing temporal variation of test environments by spatial variation of test environments  (more locations less years), (iii) moderate but critical consideration of new technologies / inovations for breeding as long as they are under development / validation in applied breeding populations following the paradigm that the best breeding program design depends on the working environment  

I think the most progress we made this year was the adoption of the Breeding Management System (BMS) as an important tool which will help us for developing a more efficient breeding program. In addition, we use an approach of MAS in the whole process of development of breeding lines and seed genetic purity control as well.

The principal tools that  I have susscessfully integrated into the breeding program are the creation of common bean differential cultivars to study the virulence diversity and evolution of the highly variable pathgens that cause rust, anthracnose (ANT), angular leaf spot (ALS) and halo blight.These differntial cultivar using Andean and Mesoamerican common bean not only essemble th genetic diversity of the common bean crop but also have been instrumental in the discovery that the virulence diversity of the rust, ANT, and ALS pathogens is a mirror imagine of the diveristy of the host. Like in the common bean, these pathogens also have Andean and Middle American strains (known as races).These studies also lead to the discovery that gnes for reistance from Andean Adean beans are very effective in the managemnet of race of these pathogen of Middle American origin. Similarly, Disease reistance genes of Middle American origin confervery effective resistance toAndean races f these pathogens. We have developed bean cultivars combining Andeanand Middle American disease genes for resistance.Under expereimental conditons these cultivars hve broad reistanceto rust to all known races of the rust Pathogen and under field conditions in Africa and the Americas. More recently, we began using bulk segregant analysis and hign thoughput genotypin with a SNP chips with >5,000 SNPS to discover DNA markers (SNPs, KASPs, SSRs) that are highly accuratein tagging rust and anthracnose disease resistance genes. One of these markers (for the Ur-3 rust resistance gene) is highly specifi and accurate and tags the gene even when the bean pant has are highly specific  other genes rust resistance genes closely linked to Ur-3. Thus the combination of phenotype based on the diversity of the bean crop and the pathogen and the integration of genomic tools has allowed us to acelerate the breeding program for reistance to highlyvariable pathogens of common bean and to make th strategy more effective and accurate., 

The breeding program in Malawi has mostly employed conventional breeding methods and participatory variety selections. Of recent, we are trying to use the breeding management system to plan our breeding program. 

Marker-assisted selection for relatively simply inherited traits has been integrated into our breeding program during the last decade.  Infrastructure and willingness to adopt new technologies led to success.

Limited amount of financial resources for breeding of minor crops and for organic and low input conditions

A) difficulty to obtain genetic material for pre-breeding from other organisations and genebanks

B) limited trials and integration of replicated on station and participatory on farm variety trials representative for given pedoclimatic condition and farming systems

C) exchange with other breeders on global level for minor crops

C: Our research follows problem identification by local farmers and also urban gardeners; building local capacity and finding funds to follow up on problems identified is often difficult.

Limited funding is the main constraint which has direct implication on human resources and expaning the scale of breeding activities.

A. Germplasm characterization - funding, facilities/protocols for phenotyping certain traits; restirictions on germplasm movement

B. Variety testing - funding to scale the testing (number of varieties, replications, locations etc), limited representative test sites in some countries, limited resources capacity of programs in some countries to handle multi-location testing 

just answering C) for the moment. Germplasm exchange is a massive constraint, especially between countries. I believe it's a critical issue for food security. 

with advances in genetics/genomics, characterization has become much more straightforward, though data is often biased toward the elite cultivated material. Access to germplasm, without impeding agreements/MTAs, continue to affect what germplasm is characterized and utlized for improvment.

with advances in genetics/genomics, characterization has become much more straightforward, though data is often biased toward the elite cultivated material. Access to germplasm, without impeding agreements/MTAs, continue to affect what germplasm is characterized and utlized for improvment.

Characterization is much more straightforward for many of the crops we work with, but the access and/or MTAs attached to germplasm can prohibit characterization and/or use. The crops community should be much more forward thinking in the sharing of data for germplasm as rsources are limited.

Main constraint: Too many traits to blend in and too many market classes of beans, each one of them with their unique set of traits/requirements. Having too many types is both a blessing (diversity) and a nightmare (lack of focused effort).

Germplasm constraint: phenotyping/genotyping capabilities and in some cases access to germplasm

Variety testing constraint: Having enough multi-location field testing in your target region.

Overall: in comparison to cereals, horticultural crops, and other groups, grain legumes in general receive substantially less funding for research and development. Historically, we have given more importance to energy (calories) and high value, than to protein. Still grain legumes are expected to advance at the same rate as others, which it has been accomplished in some cases, but with much more slower/small gains over time. It is also imporatnt to note that yield gains have been more difficult in legumes because most of the times farmers will grow cereals in their good land and put the legumes in their poor/marginal soil.

Access to germplasm

Ability to conduct large scale phenotyping trials at multiple locations

Long term constraint 1: availability to the full suite of germplasm found in all of the bean programs in the world.  Restrictions that require tracking of alleles from that germplasm as it moves from program to program have to be much less burdening.  Long term constraint 2: a full characerization of the haplotypes in all breeding programs.  This knowledge can drive parental selection.  Long term constraint 3: a stable funding level that assured each year that is provided by local and international sources.  This level can always be augmented as new programs are designed and additional funds come available.  This will ensure that a base level of population development, screening, and selection is not disrupted.

a) germplasm characterization - lack of facilities (irrigation, coldroom)

b) Variety testing - lack of qualified personnel (agronomists...)

c) Others - lack of funding in order to get more people into the program and even acquire vehicles in order to reach more farmers

Very few centers have free germplasm.Centers in the main diversity areas are unable to share due to treaties such as the BDT, Nagoya Protocol and national laws. International centers that have the mandate to maintain, understand and share germplasm with the whole scientific community tend also to run into issues. Piles of paperwork, lawyers and agreements are needed to obtain materials from other coutries. This is a huge challenge for the community, in which the crop has such a narrow genetic base and, consequently relies on foreign materials to enrich allele diversity. 

New breeders properly trained - the know how to breed (aggregating activities into one target which is the better variety for a region / sub-region and to come with this within 4 years (at least within 3-4 year potential candidate varieties / advanced breeding clones better than check must be visible). Knowledge how to breed is rarely found attribute among the next generation whereas for single innovations many specialist are available.  

We've stablished a quite good molecular lab. and that has enabled us to routinely use a marker asssisted selection approach for developing new mapping populations, characterizing the germplasm we're using in our breeding programs as well as in the control the seed genetic purity control but we just establish a biochemistry lab. and is not yet well equiped. For instance, we started working on oil qualiy but we'rev just using molecular markers. A NIRS will be very helpful for us for identifying genotypes with high oleic content at the stage of the segregating populations. In a near future, we would like to characterize our germplasm for other marketing quality traits and lack of such equipment may hampered our attempts. Our intent is also to have three generations per a year in terms of populations' advancement but we haven't glasshouses.

Human resources (qualified technicians and research assistants) are also very limited in our breeding programs and most of NARS are facing the same situation. Financial support for hiring and training technicians and young scientists is a key need for all breeding programs of NARS.

The main constrain, by far, to my program is funds to do research. The constrains in germaplasm is the narrow genetic base of the common crop and the broad spectrum genetic base of certain common bean pathogens.  Also the under utilization of the Andean gene pool in breeding beans of Middle American origin.  In regard to germplasm characterization - when searching for novel disese resistance genes,it is dififficut to phenotype wild germaplsm for disase resistance.Thus, therearevery few gnes for reistance from wild beans.Also there is a sever lack of isolates of important pathogens form wild beans.The broader gentic diversity of wild beans suggest that wild beans arealso under utilized in disease resistance.

The main constraint has been funding (adequate and timely) to support the breeding programme. IN terms of Germplasm characterization - Difficult to access and collect garmplasm from elsewhere because of restrictions and also funding.

In term of variety testing -  done in a few sites and once in a year because of lack of irrigation facilities and inadequate human personnel to oversee activities involving variety testing.

However, in general, the greatest challenge has been high rate of staff turnover (breeders and technicians) because of retirements, leaving for greener pastures - as a result it has been difficult to make and implement long term breeding plans.  This is also linked to inadequate funding and lack of incentives which somehow demotivates the spirit of hardworking among staff.

it will be very difficult to obtain genetic material due to extended property rights and patenting

huge dependency on few global players in the seed market that control major agricultural crops and loss of small and medium sized breeding organisations as well as loss of public breeding programs.

increased loss of transparency of breeding methods applied

new models are needed to finance breeding of locally important crops 

There is still massive prejudice against organic solutions from much of the agricultural research community, and a lot of dis-information from agro-chemical companies, although some (such as BASF) are starting to work with us to identify possible biological products and solutions.  In five years time, we hope that the problem areas will have shifted to actually carrying out the research, which is so critical in these times of climate change, rather than arguing that it is needed!

I do not think there will be a big difference between today and in 5 years time in terms of challenges in increasing genetic gains in the developing world. Breeding programs need to work towards improving effectiveness and efficiency in dveloping improved varieties so that varieties turn over speed is improved in the face of changing climate. Currently varieties turn over is very slow, particularly in SSA, where farmers grow very old varieties (more than 15 years even 50 years old). Phenotyping tools/facilities, human capcities/skills, funding ... etc remain challenges on the way to make progress in effectiveness and efficiency

The issue of climate change will become more problematic and we just cannot predict the effect of those changes. Climate change may mean less rainfall or the opposite in some cases, so selecting for a moving target is next to impossible.  The expectation that yields can be increased as climatic conditions change for the worse is false and is not being addressed.

Access to germplasm will get harder and harder. Climate change is difficult to address from a breeding standpoint because is so variable/unpredictable and plant breeding is a long-term effort rather than a short-term response to drastic changes. Reduced funding for research may also be a challenge. Finally, finding enough professionals interested in field/applied plant breeding.

The greatest challenge is access to assemble a team with personnel trained in the wide range of knowledge and skills needed to advance breeding of these more recalcitrant crops such as such sweetpotato that is polyploid, vegetatively propagated. 

The greatest challenge five years from now will arise from not addressing climate change today.  The economics of crop production in the developed world can be considered over a period of three to five years and can withstand climate induced changes from year to year.  Productivity in developing countries must be maintainined year to year.  What germplasm base and production system will ensure that.  Research must very active in that area today so that solutions are coming on board in five years. 

I agree with many here - basic researchers. Breeders, botanists, plant pathologists. A single major concern of mine is that in some years, the main Arachis specialists will be retired, so, species identification will become much more difficult. However, traineed personell need long term jobs to carry out their research, and some times jobs can be hard to find. In developing countries, a large number of well trained PhD students graduate and end up giving up science (or moving elsewhere) after not being able to find jobs. International funding can help retain talents.

Sustained funding for strategic breeding which is comprising 5-6 relative small units under the umbrella of 1 hub by region / sub-region such as South East Asia or Sub-Sharan Africa

Senegal is sitill sharing with some others countries the market of peanut oil and peanut cake but because of the need to meet very stringent sanitary requirements something needs to be done mainly at the level of farmers to mitigate the aflatoxin contamination. An integrated control of the contamination seems to be the only solution. Then there's some ongoing efforts that need to be suported more that ever.

The capacity to make genomic and proteonomic technologies easy to use tools for breeding crops (such as beans) with less commercail value than maize, soybean, rice, wheat, others - particulrly in developing nations. 

The greatest challenge will not be different from today - funding (direct support to breeding programs) and human personel (training and recruitment of staff) to work in the breeding program.

most important is local capacity building and empowerment of smallscale farmers especially female farmers, training farmers in cultivar testing, seed multiplication, seed quality and plant breeding

Establishement of a great number of participatory breeding initiatives for a wide array of crop species 

Facilitators helping to spread information and exchange of knowlege and genetic material,  networking among small initiatives and increased involvement of the whole value chain

New models for the ownership of cultivars, no patents on living organisms, but models fostering the right of access to genetic material as well as access to water and soil.    

Yes, obstacles are often the conviction by researchers that "we know best".  Structuring rsearch processes to optimise both researcher knowledge and farmer knowledge takes sensitivity, time and resources.  Often, breeders battle to find linkages with farmer groups.

Usually there are very few breeders (1 or 2) per breeding programs for a crop in a given country. Often these breeders may not be complemented for other skills such as pathologist, entomologist, biometrician/statician, social scientists. In some cases, the breeders may not have opportunities to be exposed to recent technological advances (e.g. genomics, data analysis & interpretation approaches...etc). Exchange visits, refresher seminars, mentoring ...etc could be useful. Besides, even if breeders may be exposed to recent technological advances, breeding programs of may not often have the necessary funds to improve field and laboratory facilities for improved efficiency

Too much emphasis is given to training young scientists in newer or modern techniques. As a result when they attempt to run a breeding program the tried and proven methods of field selection in large populations eludes them as they select for a marker linked to a single gene trait. Successful varieties are an accumulation of favorable traits that do not pop up in small diverse populations but only emerge from years of field selection starting with large populations in which rigorous selection is practiced.

Personnel with quantitative genetics/genomics skill are in high demand. As are personnel with bioinformatics and genomics experience in these key crop species. Access to training at an early stage of graduate school could be one mechanism address this as well as exchange programs for more senior personnel where they are embedded in active programs to learn the requisite skills.

The most needed capacity for all public programs, in developed or developing countries, is staff that are highly fluent in data management.  It will be essential to link phenotypic and genotypic data a transparent method so selection can be in a timely manner.  The critical decisions often must be made in a few weeks, so it is important that data input is made seamlessly when possible (genotypic data) and timely (phenotypic data).  All programs need statistical expertise that will be able to extract the relevant results from which selections will be made.  This includes standard statistical calculations as well as statistical and quantitative statistis.  Genotypic data should be managed it a central location.  This will require the development of regulations that will allow biological materials to be screened to move quickly to the screening center.  

Most needed expertises: breeders, plant pathologists, plant physiologists. High tech tools are attractive enough to get young people`s interests, but not 'old-fashioned' science, that do not yield high impact publications, but is essencial for better crops.

Obstacles that prevent effective se of trainings - it depends on the country. I have witnessed several cases of well trained, enthusiastic scientists being able to do very little due to their institute`s limitations, that are some times very basic (some places are not even set up to receive and/or utilize foreign funds effecively). It is very improtant to understand what is and what is not possible to be done in a particular place and make sure that things can be done elsewhere (eg. sequencing, high throughput genotyping, certain chemical analyses). 

(i) Plant breeding statistics, management of breeding programs and resource allocation in breeding (ii) Attracting and retaining breeders / staff – high turn-over in the public sector / breeding

Technicians, research assistants, vehicles and lab. with equipments are human and instutional capacity are the main constraints.  CGIAR centers and International advanced research centers are most of the time offering a tailored training for research personel from NARS but generally the financial support may be a challenge.

What type of capacity is most needed? The capacity to understand the local environment (soil, biotic constraints and their unique diversity, water, climaticconditions, the impact of climate change, etc) and the lack of teams to design themost effective solutions. These solutions are often specific and different from developed to developing nations, from the Americas to Africa, etc. Often the most effective solution to complex problems requires not only a team  approach but also the combination of traditional technologies (agronomy, horticulture, plant pathology) and advanced technologies (including genomics, epigenetics,and bioinformatics to analyze big sets of genomic, climatic,etc. data)

The expertise needed most is at technical level - extension staff (linking farmer with the breeder) - they need to have an understanding of the breeding process, how to collect data, how to solicit information from farmers - This also applies to research assistants - they need to have technical know how of collecting data and basic principles of experimental designs, etc. Irrigation facilities and transport are also needed to effectively run trials and collect data adeqautely.

Effective use of trainings  is prevented because - of lack of facilities - for instance - the breeding section can have only 1 computer used by the breeder, technicians and all staff - to enter data, analyse and write reports. Infrastructure like glasshouses to run experiments which can easily monitored are not there.  

Training is most effective when knowledge gained is immediately applied.  While an understanding of basic concepts is needed prior to application of new information to a breeding program, designing a training to be customized to a breeders program might assist adoption.  For example, training could be most effective if a genotyping experiment were designed with prior consultation, genotyping data generated, followed by analysis of data with breeder/other experts in a face-to-face setting.

Participatory breeding approaches involving local farmers but also the processors will have greatest impact as newly developedcultivars will be immediatly be implemented in farmers field if they were inculding in the process of breeding and cultivar testing. Training on high quality of farm saved seed production and seed quality testing will reduce production costs for farmers. The number of crops cultivated per farmers need to be increased to spread risk for unforeseen weather or market events.

Easy applicable screening tools for pest and diseases and quality testing 

Techniques which help farmers to save and preserve their seeds are vital.  As Monika has said, participatory approaches are bneeded, such as the well-know "Participatory Rural Appraisal" techniques which we have worked with for 25 years.  My focus is now on comparing conventional and organic farming systems, and mapping long-term changes in soil fertility, soil organic matter, soil biology and water use efficiency, as well as comparing nutritional quality.

Complmentary tools/approaches should be worked on to make the breeding programs more effective and efficient including:1. Developing product lines; 2. widening the gene pool; 3. optimizing selection strategies; 4. speeding up early generation advancement process and improving phenotyping approaches/facilities; 5. Optimizing testing strategies (target environments, number of entries, replications, test locations...etc); 6. adopting adavanced statistical tools for better analysis of performance data; 7. exploiting advances in genomics...etc

Advances in yield breeding in soybeans has been modest when you consider the economic investment in private sector breeding. It is not the new tools that contribute to yield but the large number of breeders running large scale selection nurseries across a broad array of environments.  The financial investment in soybean breeding by private sector is huge and is focused on a single seed type, that is not directly consumed.  In contrast, bean breeders have to focus on a large number of seed types and consider end use consumer acceptance that reduce dramatically the acceptable individuals that can be advanced. In tropical regions a wide array of disease factor similarly reduces the acceptable individuals that can be advanced.  It seems that these constraints are never considered when comparisions are made.  Soybeans in the US are grown on the best land in the country whereas dry beans are grown in less desirable areas limited by latitude and altitude.  This occurs worldwide in countries like Brazil, Mexico and short rainy season of Central America. Insufficient importance is given to the crop

Tools: extensive multi-location testing of breeding material with proper phenotyping/genotyping selection tools. Marker Assisted Breeding and not "Marker Directed Breeding". Emphasis still needs to be in the actual field performance. I don't think genomic selection work well for self-pollinated crops, which is the case for most legumes. But I'm still hopeful... let someone else figure out and then we'll use it. We have bigger problems to solve in the meantime.

Approaches: This may sound a bit harsh but it is not my intention to be: one of the main difficulties I see for research in developing countries is having to learn how to deal with the local culture and in many cases bureaucracy to make things happen. It is very common to see very good trained personnel returning to their country with great enthusiasm and ideas to develop, only to find out they hit a wall at their institutions. Few of them persevere and are succesful, but others give up after a while. How to solve this? I don't have the answer....

What a question! So many essencial tools. To increase efficiency to disease resistance breeding, use of wild relatives to widen genetic pool coupled with high throughut genotyping (done by specialized centers or companies) holds great promise, as it is already delivering good, resistant lines. Many others can be cited (improved phenotyping tools for drought and for aflatoxin resistance, imprioved field early diagnostics for disease, identification of lines with high dormency etc etc) .

(i) within 5 years: Statistics and Data management – intensive iterative 2 weeks training and updating skills regularly to the breeders community and students with additional on-line (certified) training tools, rapid phenotyping in early breeding stages – (ii)  within 10 years: off-spring testing, moving into heterosis exploiting breeding schemes & reducing breeding cycles by shuttling, overcoming seed bottlenecks for RTB (in-vitro germination, rapid cleaning up, and true seed strategies for RTB), un-maned phenotyping – (iii) within 10 years  plus: genomic selection will become cheap enough to conduct first stage testing by molecular screening which can largely increase the overall test capacity of a breeding program 

The BMS and use of molecular markers will boost legume and RTB breeding programs.

It is very important that developing-country legume programs use effective genetic solutions, particulrly to solve the widespread biotic (diseases and insects) problems of common bean and other legumes in thetropis and subtropicalregions.Properly used gentic solutions are very effective,easy to adopt byproducers, cost-effective, and without rist to theenviroment. Thus, broad genetic base of the crop is essential. Plant breeders, plant pathologists,agronomists, etc., that can use a combination of traditional and advanced (High throughput genotyping) tools. Phenotypic and marker-assisted selection are critical to develop disease and insect resistant crops in the tropics.                                                         Obviously, the tools may vary acording to the crop and the location. There is a lot of private and even public research in soybean, maize,and other commercially important crops. However the problems in soybean production in the US, Brail and Africa are unique to the region. There is no soybean rust in the US compared to the severity of this disease in Brazil or Africa

Participatory variety selection (with farmers/end users/consumers) has been very effective and instrumental to development and releasing of new varieties. However after breeding there is need for an effective system which gets seed of newly developed varieties into farmers hands - for example through community seed banks.

Participatory plant breeding for adaptation to local conditions, farming systems and markets

combined improvement of genotype and farming practice

I agree with Monika, plus participatory rural appraisal (PRA).

See my response to Theme 2, Question 1

Molecular markers linked to major disease resistance traits have been successful in dry bean breeding.  Many of these markers are being refined as SNP or other physical markers.  Complex traits present the greatest challenge and the findings of most QTL studies are not being utilized as they vary with genotypes and environment and account for very limited levels of genetic variability <10%.

As said before, single-gene markers have proven being useful, especially when easy to use and repeatible/reliable across different genetic backgrounds. The development of new breeder-friendly molecular markers is not going at the same pace as the new developments in genomics and molecular biology. The race seems to be about who can do it faster, cheaper, and with more resolution but barely any of that technology is being translated into real selection tools for breeding programs. Lots of great discoveries and new info about gene location and function, but few genomicists are willing to go beyond that because is "less interesting" and rather more utilitary.

(i) Easy applicable programs for estimating trait values in large (N=10,000 clones) replicated field designs, (ii) off-spring testing and linking to breeding values for parental material, (iii) NIRS for quality screening across crops / XRF for Fe and Zn estimates.

For legume crops various genetic tools are important: appropiate screening of germaplasm under field and greenhosue conditons, co-segreation analysis,marker assisted selection, bulk segregant analysis, genomic tools to accelerate plant breeding (new gene discovery, mapping of new genes, DNA tools (GBS, SNPgenotypping, fine mapping) to discover accurate markers closely linked to genes of interest.All ofthese tools, and others, can be used to developed superior cutivars in any countries - including devloping nations.

I am answering Theme 2, questons 1 and 2 in a single thread.  This answer addresses those activities that are essential to improve home country breeding through local, regional, and international activities.  This answer is intended to be comprehensive because many of the activities build are complementary where the success of each step depends on the completion of previous steps.

Global Activity 1: Create a Modern Breeding Program Based on the Integration of Phenotypic and Genotypic Data

1. Perform deep sequencing of core genotypes related to the cultivated crop species.

 This investment would provide valuable diversity data that would be relevant to all breeding crops working on a crop.

 What is needed?  A core activity should include the development of chromosome-scale assemblies of the target species.  If the crop actually consists of subspecies, the assemblies should be developed for each subspecies.  This should be a higher level activity managed by those highly skilled in these types of activities.  Participation should be required of home countries to ensure appropriate genotypes are selected.

 2. Perform low-pass sequencing within home country breeding programs

 This investment captures the diversity that each program is working on.  Coupled with the deep sequencing data, an accurate collection of the genes/genetic blocks that the program is managing can be captured.

 What is needed? This should also be a joint activity of specific programs and those experts in low-pass sequencing.  The sequencing approach should be standardized so the results are complementary across multiple programs working on the same crop.  As with step 1, it is essential that the breeding programs drive the selection of materials that should be genotyped.  Those providing that information should be knowledgeable in both the principles of sequencing as while as the population structure of the species as it relates to their specific program.

 3. Create a “breeding based” haplotype map within each “market class” relevant to the various agro-ecosystems that the breeding program serves.

 While inter-market-class crosses are often possible, it is time consuming to recreate the narrow specifications needed for a specific market class from inter-marker-class crosses.  A specific market-class haplotype map will facilitate the capture of those specifications based on the haplotype blocks associated with specific market classes.

 What is needed?  This effort requires experts in population genomics and plant breeding.  Because different areas of the genome are responsible for the GxE that each program confronts, the map should be breeding focused and supported by data that points to specific regions of the genome that control key traits for each program.  That type of data can be generated by each program while steps 1 and 2 above are being completed.

 4. Develop “genomic selection” chips for germplasm associated with targeted to a  market-class specific agro-ecosystem.

 The haplotype map can then be used to create the “genomic selection” chip that can be adapted to specific programs based on the agro-ecosystem that the program is serving.  This investment may be modest at first to assess the utility of genomic selection for a specific program/crop.  For the efficient adaptation of this approach, it will be necessary to complete steps 1-3 outlined above.

 What is needed?  The chip should be developed over several generations and scaled up (or down) as more information is generated by each breeding program.  A commitment to multi-generational chip development is necessary.

 5. Arrange for timely delivery of genomic selection genotypic data.

 For marker-based selection to be useful, it is essential that the genotypic data be collected in a timely manner.  To ensure this occurs, it is best to work with a commercial genotyping service.  Agreements with those providers must be in place such that the data can be collected in a timely fashion.  Those agreements must specify narrow windows when breeding materials are delivered from the program and when that genotype data will be sent back to the breeding program.  Selections can then be made on the genotypic data.

 What is needed?  The primary requirement for this to become a reality is a person with good negotiating skills who can also assess the proficiency of the service provider.

 6. Apply genomic selection with periodic/timely retraining.

 The tools and approaches outlined in 1-5 above are applied to the program.  Each program should create a genomic selection retraining schedule based on experience and breadth of program.

 What is needed? Staff that is trained in the full range of modern, genomic-based breeding are required.  The number of individuals will vary depending on the size of the program, but at least one of those individuals must be able to work accurately with the phenotypic data that is being generated by the program and the genotypic data that is being sent to the program.  That individual must also understand the nature of the genetic materials in the program: where it came from the time of domestication to why specific crosses were made just last year.

 7. Scale breeding effort (# of crosses) to the resources available to each program.

 Not all programs are of the same scale.  Therefore it is necessary that those in the program understand what financial and human resources are available, and decisions regarding # of crosses, size of different generations, and advancement timelines should be based on the that resource base.

 What is needed? The leaders of program should be trained in the economics of breeding.  I am not aware that such training is actually available in the public realm, but such a course should be required for those leading local programs.

Global Activity 2: Create Data Management Capabilities for Phenotypic and  Genotypic  Data

1. Modernize data collection using digital tablets that have established records of durability and usability

Transcribing data collected in the field and the lab is a waste of the limited human resources available to a program.  This should be digitized.  Programs should be provided with digital tablets for field collection, software to store the data, and the capability to transfer the data to a central data storage facility.

What is needed?  Tablets in the market should be assessed for their durability and usability.  Various software should be considered as data storage tools for the program: Agrobase (http://www.agronomix.com/) and Integrated Breeding Platform (https://www.integratedbreeding.net/)

2. Establish a stable, high speed internet connection at central site of breeding program.

This is self-explanatory.

What is needed?  Conversations are needed between local IT personal and the breeding program to determine the best way to move data from a tablet to a server where data analysis software is located.

3. Establish a database management system that connects genotype selection with phenotypic and genotypic data.

It would be ideal to have a data management system in place where programs can house their data.  The system should also provide support to users.  The support should be in the form of people who can answer any range of questions: beginner to advanced.

What is needed?  Public programs are not utilizing the power that can be provided by data management systems.  Such a system should be created that enables programs to utilize it at whatever scale they feel is appropriate.  It should not be advertised as the global solution, but rather it should be a presented as a useful tool to maximize the output of the program

4. Training in database management, utilization, and querying.  

On-going training is necessary that develops expertise within the programs over time.

What is needed?  A centralized training effort that is useful while being engaging should be created.  The staff should be experienced in providing such support to breeders and not be generic training providers.  

the impact of participatory plant breeding needs more scientific and political recognition. The sucessful examples around the globe need to be scaled up by respective training of plant breeders and farmers.

Farmers especially female farmers need to be empowered, the reputation and financial of the farmers need to be improved to avoid that young persons move to cities. Consumers and farmers need to be linked to appreciate the work of the farmers in feeding the citiciens. Great emphasis must therefore be put on rural development and education of farmers in sustainable agriculture.

We have found in Uganda, Zambia, Tanzania and South Africa, that linking farmers to markets, and showing them how to use readily available natural resources (as ingredients for compost, mulch and biological pest & disease control) are the most important recipes for success.  Handouts of free fertilizers and hybrid/GM seeds disempower farmers and distort local agricultural economies.

Testing advanced lines provided by major international institutions is not breeding.  Young breeders need to be affliated in some fashion with successful and mature breeding programs and made to integrate with the entire process from original crossing through early generation selection, yield testing and seed multiplication.  Young breeders need to consider ways that they can adapt these methodologies to their programs. My greatest disappointment as I travel internationally is to see too few crosses being made among parental lines that likely will not produce useful progenies. Little thought is given to how to creat or build a successful new cultivar. Young breeders need to feel ownership of the materials they are developing not acting as breeding lines testers for materials generated elsewhere, because that pipeline will eventually dry up  eventually.

Practical experience with functional software for breeding

Tailored trainings are needed.

Hands on/ practical breeding courses using the new technologies can assist in enabling use of the new technologies

Training of plant breeders to teach the farmers in participatory breeding (Training the trainers)

decentralized breeding and cultivar testing on station and on farm collaborating with local communities

Supporting farmer seed exchange networks, developing technologies and training courses on seed saving and preservatiuon, supporting agro-ecology institutions, which help to make farmers less dependent on bought-in farm production inputs, and improve water use efficiency.

1. Phenotyping facilities/tools - including irrigation facilities for speeding up early generation advance; rainout shelters and instruments for physiology and drought screening studies

Facilities that help intergrating genomics tools - utilize already developed markers for traits of interest (e.g validated SNPs for groundnut); dvelop new markers for traits where no marker is yet available

Enhancing capacity of breeders & technicians - modern tools, statistics & approaches in enhancing genetic gains

One would think that food secruity issues would drive policy and political changes that would favor more investment in ag research and in the breeding of minor or specialty crops-- those crops that the private sector ignores as non profitable.  For example beans are a critical food secruity crop in Mexico, but every year Mexico imports a significant portion of the crop from N. America as they continue to grow the crop on margin land, lacking adequate rainfall to produce competitive yields and they  use their prime land for export crops and maize.   Without any governmental changes or political unrest due to food shortages, this is unlikely to change.

1- Adequate and cheap seed delivery systems in developing countries.

2- Plant disease diagnostic labs in developing countries and trained personnel

3- Intensive training in field/applied breeding personnel who will look at molecular markers as a tool rather than a goal.

(i) Strategic locations at work for breeding in developing countries, (ii) simple field and screen house facilities including proper irrigation, field data capture, in-vitro facilities, DNA extraction, appropriate sampling / linking to high through-put quality determination, C) 1 breeding hub for 3-5 breeding units.

Laboratories, glasshouses and rainout shelters

1. Genetic solutions to agricultural problems: trained breeders, geneticists (broaden genetic base of crops),agronomists (soil probems), plant pathologists, and genomic and related professionals.

2. Team approach to solve agricultural problems of food crops 

3. Emphasis on Food security. Production of enough food of major food crops to avoid spending large amounts of capital importing these foods (bean are imported - for example in Venezuela, Mexico, Brazil, Central America, Venezuela, Haiti, countries in Africa).

1. Full genotypic characterization of all target species

2. Human resource development in the area of managing breeding programs using genotypic data

3. Implementation of a modern data management system (hardware, software, connectivity) that incorporates phenotypic and genotypic data and enables statistical analyses both phenotypic and genotypic data.

Glasshouses, Equipments (used to collect data)/transport, irrigation facilities

Glasshouses, Equipments (used to collect data)/transport, irrigation facilities

To add on some of the priorities in Haiti

Legume / soil fertility management
Peanuts, common bean, Pigeon pea  / soil fertility management

Research on use of legume for the management of soil fertility in the context of legume/cereal agrosystems
Combining the three main legume species grown in Haiti (the fourth would be Cowpea) we are capable in focussing on all different altitude 0m to 2000m / water stress (humid to dryland) / soil type (weathered acidic tropical soil to very young entisols) / pH 4 to pH 9

legume/cereal, rotation/intercropping/conservation agriculture, soil carbon storage (increasing amount of organic matter in the soils) are some of our research priorities for the soil fertility program.