Fig. 1. Flow diagram of the methods implemented in the current study.

Step 1: reference sequences and taxonomic information extraction. We extracted all 196,344 reference “species hypotheses” (SHs at 98.5% similarity; SHs can be treated as fungal species) from the UNITE fungal database (version 8.2)⁵¹. Step 2: construction of reference database of phytopathogenic fungi. The 196,344 SHs were compared against the FUNGuild database (http://www.stbates.org/guilds/app.php; accessed September 2019)⁶, and those assigned to ‘probable’ and ‘highly probable’ ‘plant pathogen(s)’ were retained. This yielded a robust reference database containing 10,829 potential phytopathogenic SHs (ppSHs). Step 3: extracting data on the global distributions of phytopathogenic fungi. We used the 10,829 ppSHs to query the GlobalFungi database²⁵ and extracted the global spatial occurrence data (and associated metadata) for these fungi. This produced a dataset containing 19,669 samples with 5479 ppSHs. Step 4: Processing novel phytopathogenic fungi sequence data. We collected additional data from forest and cropland ecosystems, sampling plant shoots, plant roots, soils and the rhizosphere during 2017–2021. ITS amplicon sequences were generated and analysed from each of these samples following the methods used for GlobalFungi²⁵, yielding an additional 643 samples with 1172 ppSHs for inclusion in this study. Step 5: combining existing and novel phytopathogenic fungi distribution data. The published data extracted from GlobalFungi (Step 3) and our new data (Step 4) were merged into a single global dataset containing 20,312 samples with 5753 ppSHs. Step 6: Data analysis. Using the global dataset, we conducted a series of analyses to systematically investigate the global biogeography of phytopathogenic fungi. Briefly, this included examining general biogeography and diversity patterns, universal ecological dynamics, driving forces and predicting future changes of phytopathogenic fungal diversity and distributions in response to climate change. The map was generated in the R language. Source data are provided as a Source Data file.