How to combine pedigree and marker information into a single estimator for the calculation of relationships?

Relationship coefficients are particularly useful to improve genetic management of endangered populations. These coefficients are traditionally based on pedigree data. In case of incomplete or missing pedigree, they are replaced by coefficients calculated from molecular data when this information is available. However, genotyping a complete population for a sufficient number of markers can be impossible. The main objective of this study was therefore to develop a new method to estimate relationship by combining molecular with pedigree data. It will be useful for specific situations, where neither pedigree nor molecular data are complete. In a companion paper, the compatible coefficients were determined. In this study, based on simulations of pedigree and marker data, the method to combine the selected coefficients was determined. Various parameters were taken into account in the model: number and quality of the marker (e.g. marker informativeness), mutation rate, quality of the pedigree (e.g. generation-equivalents), … The combined estimator has several advantages. Especially, negative relationship values obtained in literature with molecular-based estimators in case of small inbred populations can be avoided. In conclusion, this combined estimator was originally developed for the management of an endangered horse population. It should also be a promising alternative to traditionally used estimators, e.g. for the management of small and/or rare breeds; especially in case of inbred populations, with both incomplete pedigree and partial molecular information.