https://orcid.org/0000-0002-6778-4749
Abstract
Global climate change influences the emergence, spread, and severity of rust diseases that affect crops and forests. In Korea, the rust diseases that affect Wisteria floribunda and its alternate host Corydalis incisa are rapidly spreading northwards. Through morphological, molecular, phylogenetic, and pathogenicity approaches, Neophysopella kraunhiae was identified as the causal agent, alternating between the two host plants to complete its life cycle. Using the maximum entropy model (Maxent) under shared socioeconomic pathways (SSPs), the results of this study suggest that by the 2050s, C. incisa is likely to extend its range into central Korea owing to climate shifts, whereas the distribution of W. floribunda is expected to remain unchanged nationwide. The generalized additive model revealed a significant positive correlation between the presence of C. incisa and the incidence of rust disease, highlighting the role that climate-driven expansion of this alternate host plays in the spread of N. kraunhiae. These findings highlight the profound influence of climate change on both the distribution of a specific plant and the disease a rust fungus causes, raising concerns about the potential emergence and spread of other rust pathogens with similar host dynamics.
1. Introduction
Rust diseases, increasingly emerging and causing damage to major crops and trees, are on the rise across the globe [Citation1–4]. This increase is primarily linked to climate change, as it often establishes favorable conditions for rust fungi, posing threats to food security and natural ecosystems, such as forests [Citation5–7]. However, the specific mechanisms by which climate change contributes to the emergence and spread of rust diseases remain poorly understood. This lack of understanding stems from all rust fungi being obligate biotrophs, meaning that they require a living host to complete their life cycles. Additionally, the complex life cycles of many rust species, which alternate between two unrelated host plants add to the lack of understanding of their epidemiology and dynamics [Citation8].
Wisteria spp. (Fabaceae) are woody vines that are native to East Asia and North America. They are cultivated globally as ornamental trees because of their showy inflorescences and climbing growth habits. Wisteria floribunda is a popular wisteria with many commercially available cultivars. In Korea, the vine is distributed nationwide and grown for ornamental purposes and as a natural shade tree in various public spaces, such as gardens, parks, and schools. In addition, they are planted to stabilize artificially cut slopes and reduce the risk of landslides [Citation9].
To date, few fungal diseases have been reported on W. floribunda globally. These include rust (caused by Neophysopella kraunhiae), powdery mildew (Erysiphe polygoni and Erysiphe trifoliorum), leaf spot (Cercospora muelleriana and Pseudocercospora wisteriicola), and anthracnose (Colletotrichum destructivum) [Citation10]. The rust affecting W. floribunda was initially associated with Ochropsora kraunhiae (=N. kraunhiae) in Japan [Citation11]. This fungus is a heteroecious species that alternates between two different hosts during its life cycle, forming uredinial and telial stages on Wisteria spp. Later, Hiratsuka and Kaneko [Citation12] verified Corydalis species (Papaveraceae) as a spermogonial and aecial host of Aecidium corydalinum based on a pathogenicity test. This epithet is now recognized as synonymous with N. kraunhiae.
In Korea, the emergence of rust disease on W. floribunda was first reported in September 2016. Both uredinial and telial symptoms of the disease were found on the leaves of W. floribunda in an arboretum in Jeju, Korea. By April 2019, Corydalis incisa, growing near the infected wisteria began showing signs of infection, including spermogonial and aecial symptoms. Since these initial discoveries, the rust disease has consistently affected these two species. Global warming is fueling the spread of C. incisa across Korea [Citation13,Citation14]. The expansion of this alternate host could potentially enhance the occurrence and spread of wisteria rust, particularly in the central regions of the Korean Peninsula. Therefore, a predictive model is required to assess the effects of C. incisa expansion on the occurrence and spread of its rust agent under the influence of climate change. Although existing predictive models offer valuable information on some rust species [Citation15–17], developing more refined predictive models remains vital to thoroughly comprehend their epidemiology.
The goal of this study was to identify the rust agent on W. floribunda and C. incisa based on morphological characteristics, molecular sequence data, and a pathogenicity assay. Additionally, this study aimed to investigate the potential spread of wisteria rust in Korea by simulating how climate change could affect the distribution of W. floribunda and its alternate host, C. incisa, and exploring the correlation between the climate change-driven distribution of these hosts and the occurrence of wisteria rust.
2. Materials and methods
2.1. Rust samples
Rust occurrence on W. floribunda and C. incisa was investigated through extensive field surveys conducted across Korea from 2016 to 2022. The confirmed presence of rust was charted geographically to generate a distribution map. All collected voucher specimens were housed at the Korea University Herbarium (KUS) and Kunsan National University Herbarium (KSNUH) in Korea ().
Table 1. The information on Neophysopella kraunhiae specimens used for morphological and phylogenetic analysis.
2.2. Morphological analysis
The rust symptoms and sorus structures on freshly infected leaves were observed under a dissecting microscope (M205C; Leica, Wetzlar, Germany). Specific characteristics, such as the shape and size of spores were analyzed and photographed under a differential interference contrast light microscope (Axio Imager 2; Carl Zeiss, Oberkochen, Germany). Length, width, and wall thickness measurements were obtained from 50 spores in each sample, and the results were reported as maximum, minimum, and mean values. Additionally, microscopic details, such as sorus shape and spore wall ornamentation, were observed and photographed using a scanning electron microscope (S-4800 + EDS; Hitachi, Tokyo, Japan).
2.3. Molecular phylogenetic analysis
For molecular phylogenetic analysis, genomic DNA was extracted from rust-infected leaves using a MagListo 5 M Plant Genomic DNA Extraction Kit (Bioneer, Daejeon, Korea). PCR amplification was performed for the internal transcribed spacer (ITS) and large subunit (LSU) rDNA regions using primer pair ITS5u/ITS4rust [Citation18,Citation19] and LRust1R/LRust3 [Citation18], respectively. The PCR products were purified using AccuPrep® PCR/Gel Purification Kit (Bioneer, Daejeon, Korea) and sequenced by a DNA sequencing service (Macrogen, Seoul, Korea) with the primers used for amplification. The resulting sequences were edited using DNASTAR software version 7 (Lasergen, Madison, WI, USA) and deposited in GenBank (). To compare the sequences, we performed a BLASTn search against the reference sequences of Neophysopella species available in GenBank. The ITS and LSU sequences were aligned using MAFFT version 7 [Citation20]. A phylogenetic tree was reconstructed using the maximum-likelihood method based on the Tamura-Nei model in MEGA 7 [Citation21]. The robustness of the individual branches was estimated by bootstrapping 1000 replicates.
2.4. Pathogenicity test
For the pathogenicity test, aeciospores were harvested from rust-infected leaves of C. incisa using a spore collector (Tallgrass Solutions Inc., Manhattan, NY, USA) and then placed onto 1 cm2 sections of moistened filter paper (CHM Lab, Barcelona, Spain). Each paper section contained av. 500 aeciospores and three sections were placed on the lower surfaces of three healthy leaves of W. floribunda in Petri dishes. Three non-inoculated leaves served as controls. Inoculated and non-inoculated leaves were kept in an incubator at 90% humidity and 20 °C for 3 weeks.
2.5. Distribution modeling of rust and host plants
To predict the potential distribution of N. kraunhiae, we employed the maximum entropy model (Maxent) to create distribution models for rust disease and two host plants, C. incisa and W. floribunda [Citation22]. The occurrence records of these plants in Korea were retrieved from the Global Biodiversity Information Facility (GBIF.org [July 05, 2023] GBIF Occurrence Download https://doi.org/10.15468/dl.gwfwf3 for C. incisa; GBIF.org [July 05, 2023] GBIF Occurrence Download https://doi.org/10.15468/dl.x5vcd for W. floribunda). Fourteen occurrences of N. kraunhiae were utilized for rust distribution modeling (see ).
To construct the distribution model, we used 19 bioclimatic variables [Citation23] and a 90 m digital elevation model (DEM). These bioclimatic variables were derived from monthly temperature and precipitation data supplied by the Korea Meteorological Administration (KMA) and processed using the QGIS software 3.32.2 version. Pearson’s correlation test was performed using R software (4.2.2 version; R Foundation for Statistical Computing, Vienna, Austria) to assess the presence of multicollinearity. The variables selected for each distribution model are shown in . The models were constructed using four different combinations of features (H, LQ, LQH, and LQHP), along with the selected variables. “ENMTools” package in R [Citation24] was utilized to compare each model, aiming to determine the optimal ones. The optimal models identified were evaluated using bootstrapping and 10-fold cross-validation methods. In addition, the significant and influence of variables within the models were assessed by Jackknife test. With the selected models, we projected the potential distributions for the future (2050s) under two different SSP scenarios, SSP245 and SSP585, as provided by the KMA.
Table 2. List of environmental variables using distribution models.
2.6. Statistical analysis
To assess the correlation between rust occurrence, environmental variables, and the presence of the two host plants (C. incisa and W. floribunda), we performed Pearson’s correlation analysis. This analysis used rasterized data, which were projected based on the current records of the two host plants and rust occurrence data. Pearson’s correlation analysis was executed using the “ENMTools” package in R [Citation24]. Additionally, the generalized additive model (GAM), a recognized method for delineating nonlinear relationships [Citation25], was employed to construct response curves linking the occurrence of wisteria rust with bioclimatic variables and the distribution of the two hosts. This analysis was carried out using the “mgcv” R package [Citation26].
3. Results
3.1. Morphological study
The rust symptoms on C. incisa appeared as chlorotic spots on the upper leaf surfaces, with spermogonia and aecia forming on the corresponding lower leaf surfaces (). Spermogonia were hypophyllous, densely scattered over the entire surface of the leaf, orange, round, and measured 100–250 μm in diameter (). Aecia were hypophyllous, scattered or grouped, bright orange, typical Aecidium-type with peridia, and measured 200–500 μm in diameter (). Aeciospores were subglobose to ellipsoidal, pale yellow, and measured (16.2–)17.6–20.0(–21.7) × (13.2–)14.6–17.1(–19.7) μm (mean, 18.87 × 15.88 μm), with a verrucose wall surface ().
Figure 1. Rust disease caused by Neophysopella kraunhiae on Corydalis incisa. (A) Infected leaves of C. incisa with spermogonial stage; (B) Infected leaves of C. incisa with aecial stage; (C) Spermogonia; (D) Aecia; (E) Aeciospores under a differential interference contrast (DIC) microscope; (F) Spermogonium under a scanning electron microscope (SEM); (G) Aecia under a SEM; (H) Aeciospore under a SEM.
On W. floribunda, symptoms at the uredinial stage appeared as yellowish and chlorotic spots on the upper surfaces of the infected leaves (), whereas uredinial pustules formed on the lower leaf surfaces (). Uredinia were hypophyllous, mostly scattered, bright yellow to orange, round, and measured 100–200 μm in diameter (). Urediniospores were subglobose to obovate, yellowish, with two germ pores, and measured (16.3–)18.1–21.7(–23.7) × (12.1–)13.5–15.9(–17.1) μm (mean, 19.96 × 14.72 μm), with an echinulate wall of 1–1.5 μm thick (). Teliospores were hypophyllous, epidermal, two or three-celled, oblong, and measured 28–33 × 8–12 μm, with a thin wall of 0.5–1.0 μm thick (). All morphological characteristics matched well with previous data for N. kraunhiae (spermogonial and aecial stages of A. corydalinum, and the uredinial and telial stages of O. kraunhiae) [Citation12,Citation27].
Figure 2. Rust disease caused by Neophysopella kraunhiae on Wisteria floribunda. (A) The chlorotic spots on the upper leaf surface of W. floribunda; (B) Uredinia on the lower leaf surface of W. floribunda; (C) Uredinia; (D) Urediniospores under a DIC microscope; (E) Teliospores under a DIC microscope; (F) Uredinia under a SEM; (G) Urediniospores under a SEM; (H) Wall surface of an urediniospore under a SEM.
3.2. Molecular phylogenetic study
A BLASTn search of the ITS sequences of Korean samples revealed a 100% match with two reference sequences of N. kraunhiae from W. floribunda in Japan (LC492079-80), although there was a two-nucleotide difference when compared to another N. kraunhiae sequence (LC492078). Except for one sample (KUS-F32102), the LSU sequences were identical to those of N. kraunhiae isolated from W. floribunda (LC492084-5). The KUS-F32102 sample showed a 100% match with the LC492083 sequence obtained from another isolate of N. kraunhiae. In the maximum-likelihood tree constructed using the ITS and LSU sequences (), the Korean samples clustered with the reference sequences of N. kraunhiae, supported by high bootstrap values of 99 and 100%, respectively.
Figure 3. Maximum likelihood tree of Neophysopella species based on the internal transcribed spacer (ITS) (A) and large subunit (LSU) (B) rDNA sequences. The bootstrapping values higher than 70% are given above the branches. The green box represents Neophysopella kraunhiae. The Korean specimens sequenced in the present study are shown in bold.
3.3. Pathogenicity test
In the pathogenicity test (), all leaves inoculated with aeciospores of N. kraunhiae developed chlorotic spots on the upper leaf surface of W. floribunda after 3 weeks () and immature uredinia on the corresponding underside (). Urediniospores were observed emerging from the uredinia (). Morphologically, the uredinia and urediniospores were identical to those initially found on W. floribunda in the field, thus fulfilling Koch’s postulates. The control leaves remained symptom free.
Figure 4. Pathogenicity test. (A) Inoculations of rust aeciospores harvested from Corydalis incisa onto Wisteria floribunda leaves; (B) Controls; (C,D) A leaf with chlorotic spots three week after inoculation; (E) Uredinium; (F) Echinulate and immature uredinia.
3.4. Maxent model projection
Utilizing occurrence data of N. kraunhiae and its two host plants, we assessed several candidate models and selected individual optimal models: LQHP-1.5 (with an AUC value of 0.839) for C. incisa, LQHP-2 (0.693) for W. floribunda, and LQH-2 (0.893) for N. kraunhiae. In the variable contribution analysis of LQH-2, the distribution of C. incisa emerged as the most influential factor, accounting for 58.1% of the model’s importance and contributing 56.1% to the model’s performance (). The Jackknife test further validated that it acted as an essential variable in regularized training gain (). On the other hand, the distribution of W. floribunda exhibited minimal influence, with an importance of 8.6% and a contribution of 7.1%. Among the bioclimatic variables, BIO1 (annual temperature) exhibited significant importance (28.5%) and contribution (23.5%). Moreover, the analysis highlighted that all variables, except for BIO3, had a positive correlation with the model’s response curves ().
Figure 5. Evaluation of variables within LQH-2 model for Neophysopella kraunhiae. (A) The permutation importance and contribution of environmental variables; (B) The results of the jackknife tests; (C) Response curves between the occurrence of wisteria rust and environmental variables.
Corydalis incisa is primarily distributed in the southwestern region of Korea, whereas W. floribunda is widespread across Korea (). Rust occurrence by N. kraunhiae in both hosts was noted in the southwestern region, mirroring the distribution pattern of C. incisa (). In response to climate change predictions, we conducted modeling to predict the potential distributions of the two plants and rust pathogen. The projected current distributions of both hosts and rust pathogen, as shown by the model, closely aligned with the actual current distributions (). Future (2050s) climate scenarios suggested that C. incisa is likely to expand into the central regions of Korea (), whereas the nationwide distribution of W. floribunda will remain relatively unchanged (). The future distribution of N. kraunhiae () showed a pattern like to one of C. incisa, expanding gradually northward.
Figure 6. Projection for current and future (2050s) distributions of a rust fungus (Neophysopella kraunhiae) and its two host plants (Corydalis incisa and Wisteria floribunda), utilizing the maximum entropy model (Maxent) under SSP245 and SSP585 scenarios. (A–C) Corydalis incisa; (D–F) Wisteria floribunda; (G–I) Neophysopella kraunhiae.
3.5. Statistical analysis
Pearson’s correlation analysis revealed that BIO1 showed a high positive correlation (0.76) with rust occurrences on W. floribunda, whereas BIO12 exhibited a weaker correlation (−0.06) with rust presence (). The response curves generated from the GAM for BIO1 indicated an escalation in the rust disease on W. floribunda in response to increases in annual temperature, mirroring the pattern observed in its Maxent response curve (). The correlation analysis also indicated a strong positive correlation (0.86) between the appearance of rust disease on W. floribunda and the presence of C. incisa (). However, there was a weak correlation (0.15) with the distribution of W. floribunda itself. Although there was a moderately strong correlation (0.80) between the occurrence of wisteria rust and the interaction between both host plants; this was weaker than the correlation related to the probability of C. incisa being present. This suggested that wisteria rust is more strongly associated with the presence of C. incisa than with the interaction between the two hosts. Furthermore, the GAM analysis displayed a pattern in which wisteria rust occurrence significantly increased with the increasing presence probability of C. incisa (p-value: 2e-16) (). In contrast, a weaker correlation was found between the probability of W. floribunda being present and the incidence of rust disease.
Figure 7. Statistical analysis of the interaction among rust disease (caused by Neophysopella kraunhiae), its two host plants (Corydalis incisa and Wisteria floribunda), and climate variables. (A) Pearson’s correlation matrix displaying the correlations between the rust occurrence on W. floribunda and bioclimatic variables; (B) Response curves from the Generalized Additive Model (GAM), depicting the relationship between the occurrence of wisteria rust and two variables (annual temperature and precipitation); (C) Response curves from the GAM, illustrating the interdependence between the rust occurrence on W. floribunda and the distribution of the two hosts; 95% confidence intervals indicated by colored shading.
4. Discussion
Rust (Pucciniales) is a notorious plant pathogenic group of ∼8000 species [Citation28]. These pathogens are unique in their intricate life cycles, and many rust species alternate between two unrelated plants [Citation8]. Throughout this cycle, rust fungi produce different types of spores on each host plant, facilitating their dispersal and persistence, but also complicating control efforts. One well-documented rust pathogen is Puccinia graminis, responsible for wheat stem rust, which alternates its life cycle between wheat plants (Triticum spp.) and barberry (Berberis spp.) [Citation1]. Therefore, understanding the complex life cycle, including identifying primary and alternate hosts, is essential for devising efficient control methods for harmful rust diseases. This study is the first to report the effects of the rust pathogen N. kraunhiae on W. floribunda and C. incisa in Korea. We performed pathogenicity tests to verify C. incisa as a spermogonial and aecial host for N. kraunhiae, which aligns with earlier research [Citation12].
Although the connection between specific plant distributions and the onset of rust diseases is well-established [Citation29–31], little is known about how climate-induced shifts in the distribution of a host plant, such as C. incisa influence rust outbreaks. In Korea, the distribution of C. incisa was limited to the southern islands and regions until around 2010. Our predictive models using the Maxent model under SSPs revealed that up to the 2050s, C. incisa will rapidly expand northward owing to accelerating climate change, encroaching on the central regions of Korea. However, the distribution of W. floribunda is likely to remain unchanged. The potential expansion of C. incisa is supported by our field surveys and other studies addressing the climatic impact [Citation13,Citation14].
The GAM model and statistical analysis revealed a robust correlation between the spread of C. incisa and the emergence of rust on W. floribunda. The climate-driven northward expansion of C. incisa poses a considerable threat to wisteria across Korea. These findings underscore the urgent need for further research aimed at devising targeted control strategies for wisteria rust disease, particularly in regions projected to experience significant shifts in the distribution of C. incisa due to climate change.
Remarkably, despite the worldwide distribution of W. floribunda, rust disease caused by N. kraunhiae has previously been documented only in Japan and China, affecting specific Wisteria species: W. brachybotrys and W. floribunda in Japan, and W. sinensis in China [Citation10]. This localized occurrence can be attributed to the restricted distribution of its alternate hosts (C. ambigua, C. decumbens, C. incisa, and C. pallida) in East Asia, but is also aligned with our finding that C. incisa plays a pivotal role in the emergence of wisteria rust, shedding light on the potential spread of this rust pathogen. The recent invasion and rapid spread of C. incisa across other continents make this situation a cause for concern. In the United States, a weed risk assessment flagged this plant as an emerging invader with potential risks [Citation32]. Given the ongoing disease development in Korea and our predictive analysis, there is a growing concern that the migration and spread of C. incisa could trigger a global surge in wisteria rust.
This study highlights the intricate interplay between host plants, rust pathogens, and climate change, emphasizing how climatic shifts influence plant distribution and related diseases. Research indicates that climate change may not only shift the distribution of host plants but also exacerbate the severity of diseases like rust, which poses substantial risks to both agriculture and forest ecosystems [Citation17,Citation33,Citation34]. For instance, warmer temperatures and altered precipitation patterns have been linked to increased rust incidence of outbreaks. As such, monitoring rust pathogens with similar host interactions and significant economic value in agriculture and forest ecosystems is essential [Citation7]. A deeper understanding of these interactions is essential for devising effective control strategies to counteract the adverse effects of rapid global climate change. Precise disease prediction models are invaluable and emphasize the urgent need to address climate-related challenges in disease management for future ecological and agricultural planning.
Author contributions
JSL: writing—review and editing, conceptualization, formal analysis, methodology; JHP: writing—review and editing, conceptualization, methodology; YJC: writing—review and editing, conceptualization, methodology, project administration; HDS: writing—review and editing, conceptualization, methodology.
Acknowledgments
The authors would like to acknowledge the financial support of the National Research Foundation of Korea (NRF).
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
All sequence data used in this study are available in NCBI GenBank.
Additional information
Funding
References
- Figueroa M, Hammond‐Kosack KE, Solomon PS. A review of wheat diseases—a field perspective. Mol Plant Pathol. 2018;19(6):1523–1536. doi: 10.1111/mpp.12618.
- McCook S, Vandermeer J. The big rust and the red queen: long-term perspectives on coffee rust research. Phytopathology. 2015;105(9):1164–1173. doi: 10.1094/PHYTO-04-15-0085-RVW.
- Goellner K, Loehrer M, Langenbach C, et al. Phakopsora pachyrhizi, the causal agent of Asian soybean rust. Mol Plant Pathol. 2010;11(2):169–177. doi: 10.1111/j.1364-3703.2009.00589.x.
- Debnath S, Chhetri S, Biswas S. Southern rust disease of corn–a review. Int J Curr Microbiol App Sci. 2019;8:855–862. doi: 10.20546/ijcmas.2019.811.101.
- Chaves MS, Martinelli JA, Wesp-Guterres C, et al. The importance for food security of maintaining rust resistance in wheat. Food Sec. 2013;5:157–176. doi: 10.1007/s12571-013-0248-x.
- Figueroa M, Dodds PN, Henningsen EC. Evolution of virulence in rust fungi—multiple solutions to one problem. Curr Opin Plant Biol. 2020;56:20–27. doi: 10.1016/j.pbi.2020.02.007.
- Helfer S. Rust fungi and global change. New Phytol. 2014;201(3):770–780. doi: 10.1111/nph.12570.
- Duplessis S, Lorrain C, Petre B, et al. Host adaptation and virulence in heteroecious rust fungi. Annu Rev Phytopathol. 2021;59:403–422. doi: 10.1146/annurev-phyto-020620-121149.
- Kim K. An exotic invasive liana, wisteria in Korea. International Proceedings of Chemical, Biological and Environmental Engineering (IPCBEE). Vol. 40; 2012. p. 67–71.
- Farr DF, Rossman AY. Fungal Databases, Systematic Mycology and Microbiology Laboratory, ARS, USDA; 2023 [cited 2023 Feb 8]. Available from: http://nt.ars-grin.gov/fungaldatabases/
- Dietel P. Uredineae japonicae. II. Bot Jahrb Syst Pflanzengesch Pflanzengeogr. 1900;28:281–290.
- Hiratsuka N, Kaneko S. Heteroecism of the wistaria rust, Ochropsora kraunhiae (Diet.) Dietel. Proc Jpn Acad Ser B. 1978;54:300–303. doi: 10.2183/pjab.54.300.
- Lee BY, Nam GH, Yun JH, et al. EAPDC symposium: biological indicators to monitor responses against climate change in Korea. Korean J Plant Taxon. 2010;40:202–207. doi: 10.11110/kjpt.2010.40.4.202.
- Park HC. Development and application of climate change sensitivity assessment method for plants using the species distribution models. Chuncheon: Kangwon National University; 2016.
- Berthon K, Esperon-Rodriguez M, Beaumont L, et al. Assessment and prioritisation of plant species at risk from myrtle rust (Austropuccinia psidii) under current and future climates in Australia. Biol Conserv. 2018;218:154–162. doi: 10.1016/j.biocon.2017.11.035.
- Feng L, Wang H, Ma X, et al. Modeling the current land suitability and future dynamics of global soybean cultivation under climate change scenarios. Field Crops Res. 2021;263:108069. doi: 10.1016/j.fcr.2021.108069.
- Zhang N, Liao Z, Wu S, et al. Impact of climate change on wheat security through an alternate host of stripe rust. Food Energy Sec. 2022;11:e356. doi: 10.1002/fes3.356.
- Beenken L, Zoller S, Berndt R. Rust fungi on Annonaceae II: the genus Dasyspora Berk. & M.A. Curtis. Mycologia. 2012;104(3):659–681. doi: 10.3852/11-068.
- Pfunder M, Schürch S. Sequence variation and geographic distribution of pseudoflower-forming rust fungi (Uromyces pisi s. lat.) on Euphorbia cyparissias. Mycol Res. 2001;105:57–66. doi: 10.1017/S0953756200003208.
- Katoh K, Standley DM. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol. 2013;30(4):772–780. doi: 10.1093/molbev/mst010.
- Kumar S, Stecher G, Tamura K. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol Biol Evol. 2016;33(7):1870–1874. doi: 10.1093/molbev/msw054.
- Phillips SJ, Anderson RP, Schapire RE. Maximum entropy modeling of species geographic distributions. Ecol Modell. 2006;190:231–259. doi: 10.1016/j.ecolmodel.2005.03.026.
- Ramírez Villegas J, Bueno Cabrera A. Working with climate data and niche modeling: I. Creation of bioclimatic variables; 2009.
- Warren DL, Matzke NJ, Cardillo M, et al. Danlwarren/ENMTools: initial beta release; 2019.
- Hastie TJ. Generalized additive models. In: Statistical models in S. Routledge: CRC press; 2017. p. 249–307.
- Wood S, Wood MS. Package ‘mgcv’. R package version 19-0. Vol. 1; 2015. p. 729.
- Sydow P, Sydow H. Monographia Uredinearum seu specierum omnium ad hunc usque diem cognitarum descriptio et adumbratio systematica: Volumen IV: Fratres Borntraeger; 1924.
- Aime M, McTaggart A. A higher-rank classification for rust fungi, with notes on genera. Fungal Syst Evol. 2021;7:21–47. doi: 10.3114/fuse.2021.07.02.
- Kearns HS, Jacobi WR, Burns KS, et al. Distribution of Ribes, an alternate host of white pine blister rust, in Colorado and Wyoming. J Torrey Bot Soc. 2008;135:423–437. doi: 10.3159/07-RA-055.1.
- Jacobi W, Geils B, Taylor J, et al. Predicting the incidence of comandra blister rust on lodgepole pine: site, stand, and alternate-host influences. Phytopathology. 1993;83:630–637. doi: 10.1094/Phyto-83-630.
- Peterson P, Leonard K, Miller J, et al. Prevalence and distribution of common barberry, the alternate host of Puccinia graminis, in Minnesota. Plant Dis. 2005;89(2):159–163. doi: 10.1094/PD-89-0159.
- Koop A. Weed risk assessment for Corydalis incisa (Thunb.) Pers. (Papaveraceae)–Incised fumewort. United States Department of Agriculture (USDA), Animal and Plant Health Inspection Service, Plant Protection and Quarantine; 2017. p. 32.
- Zhao J, Wang M, Chen X, et al. Role of alternate hosts in epidemiology and pathogen variation of cereal rusts. Annu Rev Phytopathol. 2016;54:207–228. doi: 10.1146/annurev-phyto-080615-095851.
- Ramirez‐Cabral NYZ, Kumar L, Shabani F. Global risk levels for corn rusts (Puccinia sorghi and Puccinia polysora) under climate change projections. J Phytopathol. 2017;165:563–574. doi: 10.1111/jph.12593.