Abstract
Despite the differences between the two cultivated species Coffea arabica and C. canephora, in terms of breeding system and genetic diversity, there are some key elements in common for creating C. arabica and C. canephora varieties adapted to global warming (GW). The main existing farming systems are common to both species, and impose, for each farming system, specifications to the breeders that are quite similar for these species. To speed-up the selection of C. canephora, we propose a crossbreeding program using the widest possible genetic diversity including very closely related species (C. congensis and C. brevipes), followed by clonal selection and assisted by multi-trait and multi-environment genomic-phenomic selection. Since C. arabica contains an entire sub-genome of C. canephora, this makes the latter, the cornerstone of C. arabica breeding. For this species with very low polymorphism, the use of specific hybridisation for the introgression of resistance and adaptation genes is an absolute necessity. Introgressed materials can then be crossed with wild Ethiopian male-sterile genotypes to create high-yielding F1 hybrid. However, given the length of selection cycles, the use of genome-editing techniques enabling the transfer of large and precise quantities of DNA seems essential for introgressing several traits at once. Most of the main target traits seeking sustainability and resilience for the main farming systems are currently considered, but some constitute “blind spots”. However, for both producing species, breeding based on performance, high yield and yield stability in environments representing targeted stress scenarios, remains the proven method for selecting GW-adapted coffee varieties in the next decade.