Sorghum grain yield is highly reduced by birds in most areas in Africa where these avian pests are abundant. After much consideration on the physiological maturity of the sorghum grain, we designed an approach of early harvest of sorghum grain. The late grain soft dough was identified as the suitable stage to harvest white sorghum as it turns cream white. As the results showed it was possible to salvage over 90% of the grain yield even when high bird infestations were observed in the production plots. A test on seed germination rate showed over 80% grain viability after sun-drying the grain in the open in three different study sites. These results give hope and provide relief on sorghum production where birds are a menace.

Pangenomics has been promoted to accelerate breeding of orphan crops, but smallholder farmers in developing nations have seen little benefit so far. To address this gap, we built a global pangenomic breeding network, integrating African breeding programs, U.S. land grant universities, and international nonprofit research organizations. Here we demonstrate that pangenomics, when integrated with local crop improvement knowledge and global scientific partnerships, can facilitate breeding of drought and pest resilient varieties for smallholders. To breed trait-enhanced sorghum varieties with lgs1-1 resistance to witchweed (Striga hermonthica) for smallholders in Niger, one of the world's least developed nations, we used population genomics across local and global scales to develop lgs1-1 Striga resistance markers, and deployed them for rapid introgression of resistance into locally-preferred varieties. Genomic characterization, along with controlled experiments in laboratory, pot, field stations, and smallholder farms, confirmed lgs1-1 resistance was introgressed without loss of essential local-preference traits. New pangenomic resources, including global resequencing and graph pangenomes, further accelerated design of broadly-applicable markers. Unlocking the potential of pangenomics for stress-resilience breeding depended on stakeholder input, strong inference, South-led decision support software, and a dense collaborative network. The experience of the network provides a scalable roadmap for collaborative pangenomic breeding of trait-enhanced varieties for the world's lowest-resourced farmers.