Pearl millet genomic vulnerability to climate change in West Africa highlights the need for regional collaboration

Climate change poses a significant threat to global food security by reducing crop productivity and increasing harvest uncertainty. Sub-Saharan Africa, heavily reliant on rainfed agriculture, is particularly vulnerable. Past climate change has already reduced yields of major crops by up to 10%. Adaptation strategies are crucial, and utilizing existing crop diversity offers a promising short-term solution. This approach involves identifying currently cultivated, climate-adapted varieties suitable for future conditions, mirroring assisted migration in natural ecosystems. Ecological niche models, while helpful, often overlook intra-specific diversity and local adaptation. This study combines genomic diversity with environmental data to improve predictions of climate change impacts on pearl millet, a staple crop in West Africa, considering its adaptive potential. Pearl millet exhibits high diversity in agro-morphological and adaptive traits such as flowering time, photoperiod sensitivity, and drought tolerance, making it suitable for a wide range of climate conditions. The goal is to identify regions most at risk under future climate scenarios and assess the potential of local varieties to mitigate these risks, ultimately highlighting the need for regional collaboration in climate adaptation strategies.

Previous research has highlighted the negative impacts of climate change on crop yields, particularly in sub-Saharan Africa. Studies have shown yield reductions of up to 10% due to historical global warming. Various adaptation strategies have been proposed, including cultivating better-adapted varieties and diversifying production systems. Utilizing existing varietal diversity is seen as an efficient short-term strategy, similar to assisted migration techniques used in natural ecosystems. Ecological niche modeling has been employed to predict the impact of climate change on crop distribution, but these methods often lack consideration of intra-specific diversity. Recent advancements combining genomic diversity with environmental data provide more accurate predictions by accounting for the adaptive potential of species. These studies have shown the value of incorporating genetic information into climate change vulnerability assessments.

This study used a landscape genomics approach to analyze the impact of climate change on pearl millet cultivation in West Africa. The researchers collected genomic data from 173 landraces across 10 Sahelian countries. Pool-sequencing was employed to estimate allele frequencies for each landrace at 138,948 polymorphic single-nucleotide polymorphisms (SNPs). Principal component analysis (PCA) revealed that genomic diversity reflected the geographical origin of the varieties. A gradient forest (GF) approach was used to model variation in allelic frequencies along environmental gradients, using both genomic and climate data. The climate dataset included 157 agronomically important metrics derived from 17 climate models representing various scenarios of greenhouse gas emissions (RCP2.6 and RCP8.5). The GF models identified key climate predictors associated with genomic composition. Genomic vulnerability was calculated as the Euclidean distance between current and future genomic compositions, predicted using climate projections for 2050 and 2100. To validate the role of flowering time in adaptation, a genome-wide association study (GWAS) was conducted, identifying SNPs linked to flowering time. The relationship between genomic vulnerability and yield was assessed using a two-year common garden experiment. Finally, assisted migration scenarios were simulated to determine how migrating better-adapted varieties could mitigate climate change impacts.

The study found that genomic vulnerability displayed a latitudinal pattern, with higher vulnerability observed around latitudes 10° and 15°. This vulnerability was strongly associated with flowering time, with the northern limits of cultivation for both early and late flowering landraces showing the highest vulnerability. GWAS identified 103 SNPs significantly associated with flowering time, which were strongly predicted by climate variables in the GF model. The common garden experiment revealed a significant negative correlation between yield-related traits and genomic vulnerability, demonstrating that higher genomic vulnerability was associated with lower yield. Simulations of assisted migration indicated that long-distance, mainly transboundary migration of landraces could mitigate the negative impacts of climate change, but this requires substantial distances (mean 1059 km) and regional cooperation.

The findings highlight the importance of integrating genetic diversity into climate change vulnerability assessments for crops. The study's approach expands upon previous methods by incorporating intra-specific diversity and local adaptation. The strong association between genomic vulnerability and flowering time underscores the significance of this trait in pearl millet's adaptation to climate change. The predicted need for long-distance, transboundary migration emphasizes the necessity of regional collaboration for effective climate adaptation strategies. The successful linkage of genomic vulnerability to yield in the common garden experiment validates the predictive power of the approach and underscores the biological relevance for agricultural decision-making. The results suggest that leveraging genetic diversity through assisted migration can be a viable strategy, but potential challenges, such as farmer adoption of new varieties, need to be addressed through participatory breeding programs.

This study demonstrates the vulnerability of pearl millet in West Africa to climate change, particularly in the northern cultivation areas. Flowering time is a key adaptive trait. The research highlights the potential of assisted migration as a mitigation strategy, but emphasizes the critical need for regional collaboration to facilitate seed exchange and adoption of adapted varieties. Future research should focus on participatory breeding programs to ensure farmer acceptance of migrated varieties and further refine the prediction of genomic vulnerability to climate change incorporating additional factors.

The study's reliance on a limited set of climate models might affect the generalizability of its findings. While the common garden experiment validated the relationship between genomic vulnerability and yield, it might not fully capture the complexity of field conditions. The simulations of assisted migration do not account for the socio-economic factors influencing seed adoption. Further, the study focused on genomic data, while other factors affecting crop vulnerability like soil characteristics and pests were not incorporated.