How to deal with missing information when prioritizing CWR for conservation
When gathering information to prioritize CWR for conservation action, we often face the problem of missing information—for example, not all CWR have been Red Listed and the genetic relationships between crops and their wild relatives are not known for all crop gene pools. Missing information is a general problem in all steps of conservation planning and we need to make use of the available information in the best way possible, as well as taking into account that some information is missing.
Regarding the three main prioritization criteria: economic value of the related crop, utilization potential for crop improvement and threat status, missing information may include:
- Socio-economic value data for all relevant crops.
- Known or potential uses of the wild relative in crop improvement.
- Gene Pool or Taxon Group concepts for the wild relative.
- Threat status of the wild relative.
Options to help compensate or nullify the impact of the different types of missing information when prioritizing CWR:
- Socio-economic value data missing for all relevant crops. This situation is very common as only major crops or crop groups usually have this type of information. Rather than using socio-economic data, any indicator of economic value can be used instead, e.g. knowledge of the local, national or regional socio-economic value of crops (e.g. for particular nutritional qualities, local market value or cultural importance) (see an example here), the number of varieties of a crop cultivated in a country or region, the number of accessions of crops held in national or regional genebanks to give an indication of how important those crops are for the country/region (see an example here). However, not only do these indicators introduce a degree of subjectivity to the analysis, practitioners should be careful when ranking their importance with respect to other crops since direct comparisons cannot be made using different indicators (Kell et al. submitted). For instance, socio-economic data are often combined into crop groups (e.g. FAOSTAT presents data for millets, a crop group that includes Echinochloa, Eleusine, Eragrostis, Panicum, Pennisetum, Setaria crops). In this particular case, all crop genera within the crop group can be given the same socio-economic value.
- Known or potential uses of the wild relative in crop improvement. When the uses of a wild relative are unknown, the Gene Pool concept can be used as a proxy for the potential of the wild relative for crop improvement, as it reflects its genetic proximity to the crop. If the Gene Pool concept is also missing, the Taxon Group concept can be applied instead to determine the taxonomic proximity of the wild relative to the crop.
- Gene Pool or Taxon Group concepts for the wild relative. If these data is not readily available in the current main sources (the Harlan and de Wet inventory and GRIN Taxonomy for Plants), ideally literature searches should be undertaken. If this information is still missing and if you are using a scoring system, the score for missing information should distinguish between situations where there is no information on the potential utilization for crop improvement and cases where the wild relative is not of potential use for crop improvement, i.e. rather than scoring a ‘0’, a ‘not applicable’ would be more suitable. On the other hand, if you are using a serial system then you might consider either prioritizing only those species which have potential utilization for crop improvement information (Gene Pool or Taxon Group concepts and known and potential uses in crop improvement) or using the generic concept (or the TG4 definition, i.e. that all wild species in the same genus as the crop are CWR) and prioritize all species within the same genus of a crop.
- Threat status of the wild relative. If the extinction risk of a particular CWR has not been assessed either using the IUCN Categories and Criteria (IUCN 2001) or another (national) threat assessment system, then we can categorize these taxa as NE (Not Evaluated). If you are using a scoring system the score for NE should take into account that the species is a priority for Red List assessment and not necessarily a low priority for conservation action. In this way, there is no implication in terms of conservation planning. If you are using a serial system you could prioritize first based on the criteria for which you have data and then add to the prioritized list those CWR that have information on threat status and that are threatened (as suggested by Kell et al. submitted).
Other suggestions include:
- If a scoring system is being applied to prioritize the CWR checklist, calculate an 'index of scoring precision' per species in order to estimate the number of criteria that contributed to the score obtained for that particular CWR. It can be estimated by counting the number of criteria used in the scoring of that species, divided by the total number of criteria (Magos Brehm et al. 2012). This is only applicable when a scoring system with more than five criteria is used for prioritizing CWR, and it implies that all criteria are equally valuable, which is unlikely to be the case. How to use this information to obtain the final list of priority species or to assign different levels of conservation priority is the responsibility of the researcher undertaking the task.
- To decide not to use the criteria for which information for the majority of the CWR is missing in a scoring system. For instance, if only four species out of a few hundred have been Red Listed, it is not advisable to use 'threat status' as a prioritization criterion. It is suggested that for any prioritization criterion, if more than two thirds of the data for the taxa being prioritized is missing then that criterion should not be used.
Finally, it should be highlighted that CWR taxa that lack any of the data listed above should be tagged as 'high priority to obtain the missing information'.