Home | Profile | Achievement | Programmes | Projects | Staffs | Publications | Journals |
Biotech Glossary | Bioinformatics | Lab Protocol | Notes | Malaysia University |

A Report on BCCCA Cocoa Research Colloquium I:“Germplasm Characterisation Using Molecular Tools”
Martin Gilmour
Mars Confectionery, Dundee Road, Slough, SL1 4JX, United Kingdom


The BCCCA have had a long commitment to the improvement of the cocoa crop and undertook to hold a meeting of scientists involved in molecular characterisation of cocoa germplasm. The objectives of this colloquium were for interested scientists to share results, co-ordinate projects, and work towards a vision of a world-wide cocoa enhanced breeding programme. The colloquium was held at CIRAD on the 21st September 2000 and was attended by approximately 30 scientists. Presentations were made by BCCCA, CRU Trinidad, the University of Reading, USDA, CIRAD, and the ICGD, followed by open discussions on the future of gerrnplasm characterisation using molecular tools. The cofloquium participants agreed the foliowing:

  • There is a vital need for a co-ordinated global cocoa germplasm characteHsation programme.
  • Recording and open circulation of the full range of results from such a programme should follow the model of existing plant databases.
  • Type trees close to source will be identified and clearly labelled; the fate of off-types would be determined by the primary collections.
  • Microsatellite analysis will be used in the programme; the technology is ready to use immediately.
  • Additional experiments are however needed to refine the technique. to determine its limits, and to determine the rate and presence of mutations.

    These conclusions will be described in more detail and how molecular characterisation fits into a world-wide cocoa breeding programme enhanced by molecular tools will be discussed.


    The BCCCA Cocoa Research Strategy
    Currently, small farmers grow most of the world’s cocoa and research is urgently needed to help them produce good quality cocoa using sustainable growing systems. The UK cocoa and chocolate industry represented by the Biscuit, Cake, Chocolate and Confectionery Alliance (BCCCA) has a long tradition of support for cocoa research both in the UK and in producing countries. To safeguard the future of the cocoa and chocolate industry, there is a need to ensure sustainable supplies of good quality cocoa at prices which provide a worthwhile return to all those involved in the cocoa chain, from grower through to chocolate manufacturer. On a world-wide basis, the ultimate aim is to develop and make available cocoa trees suitable for each region, having the following characteristics: high yield, early bearing, resistance/tolerance to major relevant pests and diseases, cocoas of good quality and flavour, low maintenance and production costs, and environmental sustainability. In addition, there is a need to develop the knowledge and expertise to control pests and diseases that may arise in the future, in a way that fits with integrated crop management principles. Cf early, this task is beyond the resources, which the manufacturing ndustty is able to make available, and additional support must be obtained. However, two strategies are possible whereby manufacturers may be directly involved. Firstly, by providing assistance which will enable plant breeders in the producing countries to develop the improved tree stocks which are needed. Secondly, by sponsoring research into the control of pests and diseases and farming production practices in the major producing countries. Ideally, both of these objectives need to be achieved using no pesticides or by minimal usage in a system of Integrated Pest Management.

    Collecting, conservation, characterisation, cataloguing and distribution
    The single most important route for effecting improvements in the efficiency of cocoa production will be through plant breeding. Breeders need germplasm containing sufficient diversity to allow them to produce varieties with good economic characteristics, which will perform well under local environmental conditions and pest and disease pressures. Their progress will be accelerated if they have access to information on the characteristics of the material, are confident of its identity (perhaps through molecular characterisation), and if they can exchange material with other genebanks via intermediate quarantine to minimise the risk of spreading pests and diseases. Research needs are:

  • Collecting. Collecting of wild germplasni from the centre of origin/centre of diversity to enlarge the genetic base of material available to plant breeders. Major collecting initiatives should only be carried out, however, when material in a location is under threat or is believed to be of vital importance or where an unusual specific opportunity arises for a co-operative programme with a local institution. Planned collecting can, however, be supplemented by opportunistic collecting from elsewhere, including selected material from private and national collections, to ensure that as wide a range of material as possible is freely available in international genebanks. Collecting can be carried out in a more systematic way if the existing germplasm in collections has been characterised, preferably at the molecular level.
  • Conservation. This material must then be conserved in internationally recognised field genebanks or field collections (such as that at CRU Trinidad (International Cocoa Genebank, Trinidad) or at CATIE, Costa Rica. These may be complementary to each other but should desirably have a high degree of duplication. The need to exchange material with other collections is of great importance. In due time, other conservation methods, e.g. tissue culture or low temperature storage, may be possible and cost effective. The collections of cocoa trees in germplasm centres represent an important resource for the world cocoa community. These contain the outputs from collecting expeditions — both recent and those undertaken many years ago - important local selections, core genotypes used internationally in cocoa breeding programmes, and in many cases, genotypes of unknown origin. Molecular characterisation of cocoa germplasm will be a valuable tool to rationalise germplasm collections, i.e. make an inventory of maintainedgermplasm, identify off-types and duplications, and highlight gaps in the collections.
  • Characterisation. The purpose of characterisation is to establish the provenance of each clone and its biological and commercial characteristics, using a variety of proven techniques as available, e.g. morphological descriptors, isozyme analysis and RAPD markers (more recently microsatellite analysis). These need to be followed by screening against as wide a range of pests and diseases as possible as and when the necessary techniques are available. The possibility in the future of genemapping - identification of characteristics linked to specific loci on the genome -leading to rapid and extensive screening for desirable characteristics will become of increasing importance. This technique is not yet at an advanced stage for cocoa and will still require a substantial research effort to make any meaningful advance. It is obvious that unequivocal characterisation of cocoa germplasm by molecular methods is the best route to provide a base for future advanced genetic work on cocoa,
  • Cataloguing. The data arising from characterisation needs to be assembled in a computer database, readily usable in as wide a variety of locations as possible and providing facilities for updating and the incorporation of new types of data as these become available. The use of the internet needs to be considered, though hard copies may still be needed for some key users.
  • Distribution. Distribution of cocoa germplasm must be through internationally recognised intermediate quarantine to ensure that only healthy material is circulated, Under present procedures. cocoa seedlings need to be kept in intermediate quarantine for two years. Therefore new methods of quarantine, especially of speedy screening against a wide range of potentially harmful viruses, is urgently required to reduce the present quarantine time and costs. Inwards quarantine in the producing country is sometimes also required. Successful micropropagation procedures are needed to facilitate transfer of germplasm from quarantine. The incorporation of new genetic materials, fully characterised and catalogued, into national breeding programmes is the ultimate and urgent objective of this work. We are still using parental genotypes identified as potentially useful in early trials conducted in Trinidad in the 1940’s and 1950’s. A broader range of useful parents should be urgently identified. The BCCCA believe it is essential for the manufacturing industry to contribute to the general improvement of the cocoa crop by becoming involved in the sponsorship of research and arranging for its results to be communicated and applied in practice. Against a background of increased interest in the molecular characterisation of cocoa gerniplasm, the developments in available technology, and to build on our support for CRU, ICG,T, ICGD, and Reading University Intermediate Quarantine, the BCCCA undertook to hold a research colloquium to stimulate greater co-operation between researchers and play a part in developing a vision of a major cocoa enhanced breeding programme.

    Colloquium programme and format

    About 30 participants, mainly scientists from research institutions directly involved in the subject, attended the colloquium. Speakers were:

  • Tony Lass, BCCCA, Cadbury International Limited
  • Martin Gilmour, BCCCA, Mars Confectionery
  • David Butler, CRU Trinidad
  • Olivier Sounigo, CIRAD, CRU Trinidad
  • Mike Wilkinson, The University of Reading
  • Jim Saunders, USDA
  • Ange-Marie Risterucci, CIRAD
  • Caroline Ford, The University of Reading

    The chairman of the BCCCA Cocoa Research Committee introduced the topic of molecular characterisation using a real-life example of an area of confusion: several versions of SCA 6, which are phenotypically distinct from each other, appear to exist. The colloquium goals were described: an informal exchange of information, a description of the potential methodologies, identification of the areas of compatibility, a draft plan agreed of a global characterisation programme, and a clear recognition of the need for data capture and dissemination.

    The role of molecular biology techniques in a world-wide cocoa enhanced breeding programme was described, and the point was made that this colloquium was only considering the topic of molecular characterisation (future meetings may focus on other molecular biology tools and their use in cocoa improvement). Several more examples were presented of cases where there was a real need for unequivocal characterisation of cocoa germplasm. Real life examples were also described of attempts to characterise germplasm at CRU Trinidad, University of Reading, USDA Beltsville, and CIRAD Montpellier, using a variety of techniques. Finally in the presentation session, the extremely important issue of information management in cocoa characterization was discussed and suggestions were made as to how this could be incorporated into the International Cocoa Germplasm Database (ICGD).

    The remainder of the colloquium took the format of an open discussion facilitated by BCCCA in which the technology, gaps, agreed needs, and future steps were openly discussed by all participants. At the end of the day, approximately 30 points had been captured and these were subsequently sent by email to participants for their approval. The following are the agreed points from the BCCCA Research Colloquium I, grouped into themes:

    Agreed outputs grouped under themes

    (numbered as they came up during the colloquium)

    General strategy

  • 1. The colloquium participants agreed that germplasm characterisation is vital, and provides an essential platform for improved cocoa breeding.
  • 24. There was no time to plan a complete global cocoa characterisation programme using molecular tools. It was agreed that this would be done at a later date and that most activity for the near future would concentrate on cocoa germplasm located in the Americas - which conveniently includes the centres of diversity for cocoa.

    Linking collections and quarantine facilities to germplasm characterisation

  • 6. Samples taken for fingerprinting should be taken from as close to original Rsources as possible, e.g. Marper Farm in Trinidad in the case of the Pound collections.
  • 7, The principle of “type” specimens will be used in cocoa fingerprinting (as in herbarium collections). Maps and GPS co-ordinates will be used to locate these type trees unequivocally within collections. These type trees must be given very distinct, obvious labels.
  • 8. Individual sample trees must also be given durable labels. Cryopreservation will be considered as an additional insurance to field collections for germplasm conservation, Field collections remain essential for phenotype studies and evaluation for traits of economic importance.
  • 21. Curators were asked to tag each tree sampled for fingerprinting with a permanent stamped metal tag, tied on with galvanized wire.
  • The issue of how to deal with off-types wiil be taken at the local level, with information on decisions being fed into the ICGD.
  • 22. In the event of clear mislabelling, the issue of how to change the name was discussed. Should this be in the ICGD, does a committee need to be convened to decide, should the institution name be used, should a new name be used after a particular date, how should tissue-cultured plants be identified? are questions which need to be solved at a later date. The INGENIC newsletter was proposed as a way to raise mislabelling issues.
  • 23. It was agreed that a better view on the scale of mislabelling would be known next year, making it easier to answer the questions in point 22.
  • 15. When characterised, curators of the primary genebanks will be asked to exchange the characterised germplasm.
  • 16. The University of Reading Intermediate Quarantine Facility is currently the only quarantine station able to export cocoa germplasm to Africa (as stated by the Inter -African Phytosanitary Council). The Barbados facility is mainly for germplasm movement into Trinidad, but material could go to other quarantine stations in the same phytosanitary region. The USDA facility at the Sub-Tropical Research Station, Miami is undergoing renovation.

    Recommendations on molecular tools for germplasm characterisation

  • 2. Microsatellites (SSRs) are the way ahead for immediate use in the molecular characterisation of cocoa germplasm (in the past, characterisation methods have included morphology, isozymes, RAPDs, RFLPs, and AFLPs.).
  • 3. Fifteen SSRs are enough to use for fingerprinting cocoa. More SSRs may be needed for genetic diversity studies.
  • 4. We have enough SSRs that span the cocoa genome to now do effective fingerprinting studies.
  • 10. About 1500 samples per technician per year can be processed (leaves through to DNA, fingerprinting analysis, and interpretation). More samples can be processed if the DNA is extracted locally.
  • 13. The use of SSR primers allows for a variety of DNA separation and visualisation techniques so the use of radioactivity is not necessary.

    Additional experiments needed for a robust molecular approach to germplasm characterisation

  • 5. An approximate mutation rate of 300 per 10,000 plants could occur in cocoa germplasm, giving rise to errors in fingerprinting. An extra set of 6 SSRs would be needed to distinguish mislabelling from mutations.
  • 14. We will assess the pooling of DNA samples for SSR analysis to see if this can be used to increase the throughput of supposedly identical samples, e.g. the sets of 16 trees at Trinidad. Olivier Sounigo (CIRAD and CRU) volunteered to test the pooling of SSR samples to see where pooling breaks down.
  • 17. A ringtest was agreed between the major centres to test the inter-laboratory compatibility of the 15 primer pairs to be used in fingerprinting. Ringtest participants are: USDA—Miami, USDA—Beltsville, Penn State University, CIRAD, Nestle-Tours, University of Nottingham, CEPLAC and CRU-Trinidad. Claire Lanaud (CIRAD), Olivier Sounigo (CIRAD/CRU) and Jim Saunders (USDA) will decide on the choice of the 15 primer pairs. David Butler (CRU) will decide on five genetically diverse clones to be fingerprinted and will collect the leaves from Marper Farm.
  • 18. The ultimate test of a fingerprinting technique is if it can distinguish between full sibs. If the technique can then assign a value e.g., 99%, 98%, 97% identical etc., then this would be of real value. Ray Schnell (USDA Miami) volunteered to design and carry out an experiment to test this,
  • 19. It was estimated that to assess the mutation rate in cocoa clones, about 200 cloned trees from a single accession would have to be screened with 15 primers. It was suggested that the plant material being produced by Biofabrica in Bahia, Brazil would be ideal to test this. Mike Wilkinson (University of Reading) and Uilson Lopes (CEPLAC) volunteered to design and carry out an experiment to test this.

    Database management in germplasm characterisation

  • 11. Existing examples of plant germplasm databases will be assessed to see if they could be used for cocoa, e.g. CIRAD model, maize dbase, Arabidopsis dbase. Estimated resource need for database handling is one full time postgraduate.
  • 12. The cocoa database will try and store the fullest possible range of information on cocoa. There will be free, open access to all parties, with all information in the public domain.
  • 20. It was suggested that the next meeting of this group (perhaps during 2001) could discuss the interface between fingerprinting cocoa germplasm and phenotypic data. Phenotypic data on cocoa genotypes is usually available locally, but in the absence of a common set of agreed criteria, and data collected in different places, at different times, it was agreed to proceed with only collecting molecular data for the near future.

    Discussion and future prospects

    General strategy
    Given the level of interest shown in the colloquium it came as no surprise to find that the participants were in complete agreement that the overall objective of unequivocal characterisation of cocoa germplasm was an essential and worthwhile goal. In the longer term, there is a clear need for a global co-ordinated programme, but the USDA project to characterise the cocoa germplasm in the Americas represents a very good start. A future output of this work could be the definition of a ‘virtual collection” of cocoa germplasm, defined genetically, spanning the known genetic diversity of the species, but not necessarily geographically located at one site.

    Linking collections and quarantine facilities to germplasm characterisation
    The current collections of cocoa germplasm are clearly the first place to start on any world-wide cocoa characterisation programme. There was broad agreement that in characterisation work, type specimens should be identified by people who have close knowledge of the collections and specific varieties of cocoa. The simple but vital aspect of tagging these type trees should be done in a robust, durable way. The subject of how to deal with off-types generated a lot of debate, but apart from recognising that this was a major issue, the colloquium did not attempt to propose a solution. It would perhaps be appropriate to place the responsibility for dealing with off-types with the individual curators of the collections. As the cocoa research community moves towards a world-wide cocoa molecular characterisation programme, there is an expectation that primary genebanks will share results and germplasm. Although this is potentially a politically sensitive issue, it needs to be stated that it remains the expectation of the research community that there should be free exchange of results and germplasm between collections. This needs to be done in a way which takes into account the potential spread of cocoa diseases, therefore it was noted in the colloquium that The University of Reading Intermediate Quarantine facility is currently the only centre able to forward cocoa germplasm to all cocoa producing countries.

    Recommendations on molecular tools for germplasm characterisation
    It was not surprising given the high level of existing work in this field, that there was a lot of debate on the techniques currently used in cocoa molecular characterisation. It was more surprising that some clear recommendations emerged as to the way forward - the use of microsatellites, the number of microsatellites needed for fingerprinting (15), the estimated throughput, the preference to avoid radioactivity, etc.

    Additional experiments needed for a robust molecular approach to germplasm characterisation
    A feature of the colloquium was the planning of future, usually collaborative experiments, which would take the research community closer to a global cocoa characterisation programme. More experiments were planned to distinguish mutations from off-types to look at pooling samples for greater throughput (especially for collections like that in Trinidad where multiple copies of single genotypes are kept), to set up an inter-laboratory ringtest. and to determine if fingerprinting could distinguish between sibs.

    Database management in germplasm characterisation
    The colloquium debate was mainly restricted to the above issues and practical aspects of applying molecular biology techniques to cocoa germplasm characterisation. It became clear during the meeting that this was only part of the bigger picture, and that the issue of how to record, integrate, and share the voluminous data likely to be generated by molecular characterisation was a major unfilled gap in resources. This was emphasised by the desire of the colloquium participants to have the largest range of data stored and made available to the research community. A few recommendations were made such as to use existing models of plant genetic databases, to try and link the data to the ICGD, and to start recording molecular data leaving phenotypic data for the future. The colloquium concluded that the need for a specialised bioinformatics person was clear, and the need for funding of such a person was a priority.

    Molecular characterisation as part of a world-wide cocoa enhanced breeding programme
    The BCCCA sees the molecular characterisation of cocoa germplasm as part of a larger programme to modernise the cocoa crop. Secure, diverse, characterised collections of cocoa germplasm form a platform for breeders, agronomists, pathologists, ecologists, physiologists, molecular biologists, and biotechnologists to work on known accessions of cocoa, safe in the knowledge that their results will be internationally comparable. This work also underpins the setting up of mapping populations for the identification of QTLs, and the development of marker-assisted selection schemes. Gene discovery programmes using EST technology, or candidate gene approaches also rely on the unequivocal identification of germplasm. These modern molecular breeding techniques and molecular characterisation in combination with conventional breeding, will lead to the development of elite trees which can be propagated by a variety of new tissue culture and horticultural methods. It is not an exaggeration to say that the outputs from this colloquium are at the heart of improving the cocoa crop for growers and consumers around the world. Therefore in conclusion, as well as the above outputs and action points, the unequivocal message from the first BCCCA Research Colloquium is that molecular characterisation has a vital role to play in a world-wide cocoa breeding programme enhanced by molecular tools.


    The BCCCA would like to thank the speakers, the participants, and our hosts CIRAD for making the first BCCCA research colloquium a success.