Published online September 14, 2017 RESEARCH Phylogenetic Prediction of Alternaria Leaf Blight Resistance in Wild and Cultivated Species of Carrots Carlos I. Arbizu, Pamela M. Tas, Philipp W. Simon, and David M. Spooner* C.I. Arbizu, Dep. of Horticulture, Univ. of Wisconsin, 1575 Linden ABSTRACT Dr., Madison, WI 53706-1590; C.I. Arbizu, current address, Instituto Plant scientists make inferences and predic- de Biotenología, Facultad de Agronomía, Univ. Nacional Agraria tions from phylogenetic trees to solve scientific la Molina, Av. La Molina s/n, Lima 12, Lima, Perú; P.M. Tas, Dep. problems. Crop losses due to disease damage of Chemistry, Univ. of Wisconsin-Platteville, 1 University Plaza, is an important problem that many plant breed- Platteville, WI 53818-3099; P.W. Simon and D.M. Spooner, USDA- ers would like to solve, so the ability to predict ARS, Vegetable Crops Research Unit, Dep. Horticulture, Univ. of traits like disease resistance from phylogenetic Wisconsin, 1575 Linden Dr., Madison, WI 53706-1590. Received trees derived from diverse germplasm would 6 Feb. 2017. Accepted 4 May 2017. *Corresponding author (david. be a significant approach to facilitate culti- spooner@ars.usda.gov). Assigned to Associate Editor Vasu Kuraparthy. var improvement. Alternaria leaf blight (ALB) is Abbreviations: ALB, Alternaria leaf blight; AIC, Akaike information among the most devastating diseases of car- criterion; AUDPC, area under the disease progress curve; EBN, rots (Daucus spp., Apiaceae) worldwide. Thus, endosperm balance number; HSD, honestly significant difference; ML, new approaches to identify resistant germ- maximum likelihood; PTP test, permutation tail probability test; PVY, plasm to this disease are needed. In a study of Potato virus Y. 106 accessions of wild and cultivated Daucus and related genera, we determined plant height Feeding a constantly increasing world population of 9.1 billion is the best explanatory variable to predict ALB in year 2050, which is ~30% above today’s population, will resistance using a phylogenetic linear regres- sion model. Using the estimated area under the require increasing food production by 70% (FAO, 2009). This disease progress curve, the most resistant spe- challenge can be addressed by employing plant breeding, with cies to ALB were the non-carrot relative Ammi disease resistance breeding as a crucial component (Miedaner and visnaga (L.) Lam. and the wild carrot relative D. Korzun, 2012). Alternative approaches for resistance breeding crinitus Desf. A permutation tail probability test are needed, since there are increasing numbers of infectious crop was conducted considering phylogenetic signal diseases caused by fungi and oomycetes (Gawehns et al., 2013). to evaluate the strength of association between Plant breeders have used germplasm resources for breeding based the Daucus phylogeny and ALB resistance. We on assumptions and observations that they possess resistances to found that species belonging to clade A, which a variety of diseases. Therefore, germplasm characterization is includes carrots and other Daucus possessing of great importance to identify accessions with sufficient levels 2n = 18, 20, or 22 chromosomes, are slightly of disease resistance to improve yields, because resistance genes more resistant to ALB than members of other can be introgressed from wild species into elite varieties ( Jansky, clades of the Daucus phylogeny. 2000; Gawehns et al., 2013; Piquerez et al., 2014). A common method to identify disease-resistant accessions in germplasm collections is to screen genebank accessions exposed to the pathogen of interest in a greenhouse or field trial. However, Published in Crop Sci. 57:2645–2653 (2017). doi: 10.2135/cropsci2017.02.0078 © Crop Science Society of America | 5585 Guilford Rd., Madison, WI 53711 USA All rights reserved. crop science, vol. 57, september–october 2017 www.crops.org 2645 limitations exist, such as limited labor or experimental than taxonomic and biogeographic prediction is required plots (Endresen, 2010). Other approaches, such as associa- for screening of germplasm collections. tion studies between the trait of interest and ecogeographic Cultivated carrot (Daucus carota L. subsp. sativus factors, have been used to select germplasm of interest with Hoffm.) is one of the most popular and commonly con- varying levels of success. For example, Peeters et al. (1990) sumed vegetables worldwide (Rubatsky et al., 1999). explored the association between salt tolerance and eco- Modern carrot production uses mechanical harvesters, geography to predict performance, obtaining only a weak and strong healthy foliage is an important feature for an association between these factors. Similarly, Hijmans et effective mechanical harvest. However, carrots are fre- al. (2003) used a general linear model regression analy- quently affected by foliar diseases, including leaf blights sis to determine the association between frost tolerance caused by fungi Alternaria dauci (Kühn) Groves & Skolko of wild potatoes (Solanum tuberosum L.) with taxonomic, and Cercospora carotae (Pass.) Solheim, and bacterial blight geographic, and ecogeographic factors, finding a strong caused by Xathomonas hortorum (Pammel 1895) Dowson association between frost tolerance and species, but that 1939 pv. carotae (Kendrick) Dye (Gugino et al., 2004; du temperature at the accession collection site was a weak Toit et al., 2005). Alternaria leaf blight (ALB), caused by predictor of frost tolerance. Inconsistent and weak asso- A. dauci, has spread to all carrot production areas in the ciations were also obtained between resistance to white world and is considered the most destructive disease of mold [caused by Sclerotinia sclerotiorum (Lib.) de Bary] and carrots (Rubatsky et al., 1999; Vintal et al., 1999; Farrar ecogeography of collecting sites from 34 species of wild et al., 2004), reducing yields by 40 to 60% (Ben-Noon et potato ( Jansky et al., 2006). On the other hand, Endresen al., 2001). The most common methods of control include et al. (2011) investigated predictive association between use of clean seed, crop rotation, cultivar selection, and biotic stresses and ecogeographic data for wheat (Triticum fungicide applications. Partial resistance in some carrot aestivum L.) and barley (Hordeum vulgare L.) landraces, con- cultivars offers limited protection but still requires fre- cluding that ecogeographic distribution of resistance stem quent fungicide applications (Boedo et al., 2010). There rust (caused by Puccinia graminis subsp. graminis Pers.:Pers.) has long been interest in genetic resistance and other non- in wheat and net blotch [caused by Drechslera teres (Sacc.) fungicidal approaches to control of ALB (Farrar et al., Shoemaker] in barley are associated with climatic factors. 2004). Therefore, breeding for cultivars with higher levels Another plausible strategy to facilitate the accurate of resistance to ALB, and durable resistance to A. dauci, is identification of germplasm with a trait of interest is to use of major interest for carrot breeders (Boiteux et al., 1993; taxonomic information of the species to focus on germ- Boedo et al., 2008; Simon et al., 2008). plasm more likely to bear that trait, making the process of The latest comprehensive taxonomic monograph of germplasm characterization more efficient by saving time Daucus by Sáenz Laín (1981) recognized 21 species divided and resources. Historically, it has been assumed that tax- into five sections: Anisactis DC., Chrysodaucus Thell., Daucus onomy has the ability to predict the presence of traits in a L., Meoides Lange, and Platyspermum DC. Daucus species group for which the trait has been previously characterized maintained in germplasm banks could provide additional in a representative subset of the group (Jansky et al., 2006; sources of genetic resistance for ALB that would benefit Spooner et al., 2009; Cai et al., 2011). The validation of this cultivated carrot. With this in mind, it will be useful for assumption was investigated by Jansky et al. (2008) infer- carrot breeders to identify resistant carrot genotypes by ring predictability of early blight [caused by Alternaria solani using predictivity approaches. The purpose of the present (E&M) Jones & Grout] in potato from the associations of study is (i) to evaluate resistance to ALB among Daucus taxonomic information and environmental variables, show- species, and (ii) to investigate the association between the ing that monthly average precipitation in July was the most Daucus clades and ALB scores. To this end, we employed discriminating factor to predict resistance to early blight. the phylogenetic tree generated by Arbizu et al. (2014b), In addition, Cai et al. (2011) conducted a test of taxonomic using 94 nuclear orthologs, and screened 106 accessions of and biogeographic predictivity to resistance of Potato virus Y wild and cultivated carrots. We conducted an evaluation (PVY) in wild potato germplasm, finding that wild potato of carrot vigor and ALB disease resistance in the field, species with an endosperm balance number (EBN) of one then developed a trait evolution model using the Daucus shared stronger resistances to PVY than species with differ- maximum likelihood (ML) tree (Arbizu et al., 2014b) and ent EBN values. In addition, they showed that populations a phylogenetic linear regression model. The most signifi- of wild potatoes from low elevations were more resistant cant explanatory variable was predicted by using the ALB than populations from high elevations, even though the scores. Finally, to determine the strength of association mean of the predictors had a widespread and a low pre- between ALB resistance and the classification of carrots, dictive value. Similarly, Spooner et al. (2009) studied the a permutation tail probability (PTP) test was conducted, resistance to disease and insect pests in wild relatives of cul- employing the phylogenetic signal as the criterion. tivated potato and concluded that a more effective approach 2646 www.crops.org crop science, vol. 57, september–october 2017 MATERIALS AND METHODS values to evaluate differences between species. The AUDPC Plant Material mean comparisons were conducted by using Tukey’s honestly significant difference (HSD) test with the function HSD.test in We evaluated 91 accessions of Daucus and 15 accessions of the the Agricolae package. related genera Ammi, Astrodaucus, Caucalis, Oenanthe, Orlaya, A phylogenetic linear regression was employed to predict Pseudorlaya, Rouya, Torilis, and Turgenia for a total of 106 acces- ALB scores in wild and cultivated species of carrots considering sions collected from 21 countries (Supplemental Table S1). All a ML tree obtained by Arbizu et al. (2014b), plant height, and accessions were obtained from the United States National Plant plant width. First, we fitted six phylogenetic linear regression Germplasm System, maintained at the North Central Regional models (lambda, Brownian Motion, Kappa, Ornstein-Uhlen- Plant Introduction Station in Ames, IA. Further details of the beck model with the ancestral state at the root having the accessions examined in this study are available at the Germplasm stationary distribution, Ornstein-Uhlenbeck model with an Resources Information Network (GRIN, https://npgsweb.ars- ancestral state to be estimated at the root, and Early Burst) grin.gov/gringlobal/search.aspx). using the phylolm function in the phylolm package (Ho and Ané, 2014) in R, with only ALB scores recorded 93 d after sowing Field Experimental Design (harvest time) as the dependent variable, and plant height and and Disease Phenotyping plant width as predictors (i.e., explanatory variables). We then All 106 accessions were direct seeded by hand in 1-m ´ 3-m performed model selection based on the Akaike information observation plots at the University of Wisconsin Hancock criterion (AIC). That is, the model with the lowest AIC value Agricultural Research Station in Hancock, WI, with two repli- was chosen. A backward stepwise model selection for phyloge- cations per accession. Natural ALB infestation occurs on carrots netic linear model was conducted using the function phylostep in in this research station; therefore, plots were not artificially the phylolm package. Finally, since we were interested in deter- inoculated. Nitrogen fertilizer was applied at the beginning mining whether certain clades of our ML tree contain entries and middle of the growing season. Plots were weeded by hand that possess higher levels of ALB resistance, a comparison and hoe, and plants were hand thinned to 5 to 6 cm in the row, among Daucus clades (Fig. 1) was performed as follows: (i) clade leaving approximately 50 to 60 carrots in each plot. An ALB A vs. clade B + outgroup, (ii) clade B vs. clade A + outgroup, susceptible cultivar, ‘Heritage’, was used as infection plots that (iii) clade A vs. clade B, and (iv) subclade A¢ vs. the remaining are reliably attacked by naturally occurring populations of A. species in clade A + clade B + outgroup. Comparisons were dauci. In addition, ‘Bolero’, a less susceptible cultivar, was also performed using a PTP test to determine whether our data con- included as a control. Presence of A. dauci conidia was confirmed tain phylogenetic structure (Baum and Smith, 2013). The ALB by Tas (2016) following the protocol described by Strandberg scores were permuted 1000 times using R, randomly assigning (1983). Subjective ALB ratings were scored for each accession states to taxa. Permuted datasets were subjected to phyloge- plot by examining leaves in 10 random sites within each plot six netic signal estimation using the phylolm function. Phylogenetic times over the entire growing season using the following scale: signal was then visualized using the function hist in R. If phy- 0 = no visible disease damage, 1 = up to 25% disease damage, 2 logenetic signal for the original dataset (not permuted), which = 26 to 50%, 3 = 51 to 75%, 4 = >75%. Plant height and plant is Pagel’s Ʌ = 0.98, is higher than all of the permuted datasets, width characters were scored in the field 50 d after sowing by the original dataset can be said to have significant phylogenetic measuring 10 plants per plot, capturing the normal range of signal, determining differences among the clades that were variation. Briefly, plant height was scored using a ruler from compared. Then, for those comparisons that were significant, the base of the plant to the highest point. Leaves that snapped an ANOVA of the ALB scores recorded 93 d after sowing (har- at their base or were near the ground were not considered to vest time) was performed, grouping the accessions according measure plant width; measurements were scored at the widest to the clade to which they belonged (Fig. 1). Briefly, Daucus is part. All evaluations were recorded by the same individual. contained within two main clades, A and B, and within clade A, there is a subclade named A¢ comprising the subspecies of Statistical Analysis D. carota, D. syrticus Murb., and D. sahariensis Murb., all with We analyzed our disease scores with R version 3.3.1 (R Core 2n = 18 chromosomes (Fig. 1). Finally, ALB mean comparisons Team, 2016). Means were calculated using the ddply function among clades were conducted by using Tukey’s HSD test with in the plyr package (Wickman, 2011). A descriptive statistical the function HSD.test in the Agricolae package. analysis was conducted to verify the mean, median, standard deviation, and range of values. Box plots were used to visu- RESULTS alize comparisons across accessions and to check for outliers Alternaria Leaf Blight Screening that may represent erroneous entries. Accessions were classi- Severity of ALB damage was visually confirmed, as well fied into their corresponding species names according to the as the presence of A. dauci fungi in the research plot by most recent molecular and morphological studies (Arbizu et al., microscopic evaluation by Tas (2016). Supplemental Table 2014a, 2014b, 2016a, 2016b; Spooner et al., 2014). A quantita- S1 lists the disease scores recorded for each accession. At tive summary of ALB intensity over the growing season was harvest time, complete resistance (disease score = 0) was determined with the area under the disease progress curve (AUDPC). Values of AUDPC were calculated using the audpc observed in all four accessions of species D. crinitus. In function in the Agricolae package version 1.2-4 (de Mendiburu, addition, 11 accessions of Daucus, and two related species 2016) in R. We also conducted an ANOVA of the AUDPC [Ammi visnaga (L.) Lam. and Torilis arvensis (Huds.) Link] crop science, vol. 57, september–october 2017 www.crops.org 2647 Fig. 1. Phylogeny of Daucus obtained from Arbizu et al. (2014b) using a maximum likelihood analysis based on 94 nuclear orthologs and 107 accessions. Species identities have been corrected in the Daucus carota complex and D. syrticus in subclade A¢, Rouya polygama in clade A, and members of the D. guttatus complex in clade B, according to Arbizu et al. (2014a, 2016a, 2016b). showed partial resistance (disease score = 2). Thirty-four ~10 wk), 75 accessions had ALB symptoms on <50% of accessions had ALB symptoms on >75% of the foliar area the foliar area, indicating partial resistance. However, at (disease score = 4), demonstrating very low levels of resis- harvest time (93 d after sowing), they were predominantly tance against A. dauci. Further, 77 d after sowing (i.e., scored with a value of 3 or 4, showing a progress on the 2648 www.crops.org crop science, vol. 57, september–october 2017 spread of ALB. Variation among accessions for ALB score be included in the model. Supplemental Table S2 lists the for two subspecies of D. carota was visualized using box predicted ALB scores, using plant height as the predictor plots (Fig. 2). Two members of the D. carota complex, (i.e., explanatory variable). Predicted ALB scores ranged subsp. capillifolius and subsp. sativus, illustrate the interac- from 3.1 to 4.0, indicating that no partial resistance was cession variation (Fig. 2) that exists in our dataset. Species observed or >50% of disease damage was present in acces- were ranked according to AUDPC value (Table 1) to sions of Daucus and related genera. determine the resistance to ALB. Two species, Amni vis- Comparison among Daucus clades (Materials and naga and D. crinitus, showed very high levels of resistance; Methods) to determine the presence of ALB resistance on the other hand, D. littoralis Sibthorp & Smith and Pseu- revealed that only two comparisons were significantly dorlaya pumila (L.) Grande possess the lowest significant different, with the phylogenetic signal of the permuted levels of resistance to ALB. Among the D. carota complex dataset consistently lower than the phylogenetic signal (Fig. 3), subsp. capillifolius and subsp. maximus exhibited the with the original dataset: (i) clade A vs. clade B + out- highest and subsp. sativus the lowest significant levels of group, and (ii) clade A vs. clade B. Similarly, a multiple susceptibility (Fig. 3). Daucus syrticus, which is the closest comparison procedure (Tukey’s HSD, significance level species to the subspecies of the D. carota complex (Arbizu = 0.1) showed that entities of clade A are slightly more et al., 2016a), had a higher AUDPC value compared with resistant to ALB when it is compared with members that those subspecies (Fig. 3). belong to (i) clade B + outgroup (p = 0.08), and (ii) clade B (p = 0.03). As mentioned above, D. crinitus, which Phylogenetic Linear Regression Analyses belongs to clade A, has the highest resistance to ALB. To The six models of trait evolution tested in the present determine if the significant difference found within clade study had AIC values ranging from 221.5 to 227.5. The A vs. clade B + outgroup and vs. clade B is not influenced Ornstein-Uhlenbeck model with an ancestral state to be only by the very high resistance of D. crinitus, we excluded estimated at the root (OUfixed) had the lowest AIC value. Therefore, we continued our stepwise model selection for Table 1. Area under the disease progress curve (AUDPC) in phylogenetic linear model with OUfixed, obtaining plant ascending order for 25 species and five subspecies. height as the only significant explanatory variable (AIC Species AUDPC of Alternaria leaf blight† = 220.4); plant width was not considered as significant to Ammi visnaga 4f Daucus crinitus 6.69f Astrodaucus littoralis 37.75ef Daucus glochidiatus 49ef Daucus carota subsp. capillifolius 55.75ef Orlaya daucorlaya 57ef Daucus carota subsp. maximus 58.25ef Torilis arvensis 59ef Oenanthe virgata 60.5ef Daucus carota subsp. carota 62.88ef Caucalis platycarpos 67def Daucus guttatus 70.46de Daucus carota subsp. gummifer 70.66de Torilis nodosa 72.42cde Daucus carota hybrid 72.5cde Daucus carota subsp. sativus 78.97cde Daucus pusillus 82.42bcde Rouya polygama 88.25bcde Torilis leptophylla 88.75bcde Daucus syrticus 88.79bcde Daucus involucratus 92.83bcde Daucus aureus 94.25bcde Daucus setulosus 97.33abcde Daucus tenuisectus 101.88abcde Daucus conchitae 108abcd Daucus muricatus 120.33abcd Orlaya daucoides 121abcd Turgenia latifolia 129.5abc Fig. 2. Box plots showing interaccession variation for Alternaria Daucus littoralis 129.88ab leaf blight scores for Daucus carota subsp. capillifolius and Pseudorlaya pumila 157.75a subsp. carota. † Values followed by the same letter were not significantly different at p = 0.05. crop science, vol. 57, september–october 2017 www.crops.org 2649 Fig. 3. Measurement of Alternaria leaf blight throughout a growing period (i.e., area under the disease progress curve [AUDPC]) of 93 d for subspecies of D. carota and D. syrticus. it from a new round of analyses using the PTP test. When resistance among members of clade A vs. clade B consid- D. crinitus is taken out of the analysis, the phylogenetic ering a significance level of 0.1 (p = 0.06). signal of the permuted dataset is still lower than the phy- logenetic signal with the original dataset. However, this DISCUSSION value without D. crinitus tends to be closer to the phy- A phylogenetic tree is a hypothesis of evolutionary histo- logenetic signal of the original dataset when comparing ries based on one or more criteria (here, nuclear orthologs). clade A vs. clade B + outgroup. In addition, the ANOVA Hypotheses of evolutionary histories have been used by test indicated that resistance to ALB of clade A is not sig- the scientific community for various purposes (Rønsted et nificantly different from other clades (p = 0.27). On the al., 2012; Baum and Smith, 2013) and have been assumed contrary, the ANOVA test showed differences for ALB to be useful to predict the presence of traits of interest 2650 www.crops.org crop science, vol. 57, september–october 2017 in a group for which the trait has been characterized in agrochemicals. In addition, plant scientists may find useful only a subset of the group ( Jansky et al., 2006; Spooner our results showing the predictor of carrot height, as stim- et al., 2009; Cai et al., 2011), such as disease resistance ulated here by the studies of Turner et al. (2016, 2017). (Khiutti et al., 2015). To date, predictivity studies have Additionally, our analysis predicted that no strong resistance been conducted in potato ( Jansky et al., 2006, 2008, (disease score = 0 or 1, Supplemental Table S2) to ALB 2009; Spooner et al., 2009; Cai et al., 2011; Chung et al., existed on carrot germplasm examined here. A plausible 2011; Limantseva et al., 2014; Khiutti et al., 2015), the explanation is the presence of a high disease pressure in the family Amaryllidaceae (Rønsted et al., 2012), the genus research field where this study was conducted. Plant height Euphorbia (Ernst et al., 2016), and barley (Endresen, 2010; has been commonly used on several statistical models as a Endresen et al., 2011). However, this is the first phylo- key variable to evaluate yield in corn (Zea mays L.; Mourtz- genetic predictivity investigation on Daucus or any other inis et al., 2013). Other studies indicate that the severity member of the Apiaceae. of a foliar disease caused by Septoria tritici Desm. shows a Calculation of the AUDPC has been widely used by relationship with date of heading and plant height in winter epidemiologists to assess quantitative resistance in many wheat (Tavella, 1978; Lovell et al., 1997). Relative to this crop cultivars (Jeger and Viljanen-Rollinson, 2001). Since study, further research is needed to determine if other top ALB is a polycyclic disease (many infection cycles in a size traits in carrots can help to predict resistance to ALB. season), it is recommended to calculate the AUDPC to sum- A detailed screening of traits from germplasm, like marize the disease scores (Fry, 1978). We here identified A. the data generated in the present study, is frequently used visnaga (2n = 20, 22) and D. crinitus (2n = 22) as the most in association mapping studies (Clotault et al., 2010; Jour- resistant species to ALB. These two species do not belong dan et al., 2015). We here employed phylogenetic signal to the primary genepool for carrot breeding, which possess as a criterion to test the strength of association because it 2n = 18. However, it is worth exploring the high antimi- is a statistical dependence among the traits of species due crobial capacity of D. crinitus against Candida albicans (C.P. to their phylogenetic relationships (Revell et al., 2008). Robin) Berkhout and Staphylococcus aureus Rosenbach (Ben- We only found significant association between clade A diabdellah et al., 2013), supporting the traditional medicinal vs. clade B + outgroup and clade A vs. clade B regarding application of this plant. As a result, we consider that, similar resistance to ALB. Similarly, Tas (2016) studied possible to the study of Camadro et al. (2008), D. crinitus should be associations between ALB severity and domestication tested to establish the feasibility of hybridization with the status, flowering habit, leaf glossiness, storage root color, cultivated carrot. On the other hand, we report that D. litto- and geographic origin and found only a slight correla- ralis and Pseudorlaya pumila are the two least resistant species. tion between increased ALB severity in purple-, red-, or Interestingly, we found that, despite the higher resistance white-colored root relative to orange and yellow roots. of Ammi visnaga, the ALB resistance level as a whole in the Studies in potato showed weak or inconsistent association outgroup clade of Daucus (Fig. 1) was not high. This may be of disease scores and taxonomy or geography ( Jansky et explained by the virulence of the pathogen A. dauci, which al., 2006, 2008, 2009; Spooner et al., 2009; Cai et al., was reported to be capable of infecting not only wild car- 2011; Chung et al., 2011; Limantseva et al., 2014; Khiutti rots, but also other wild Apiaceae (Neergaard, 1977; Soteros, et al., 2015). Perhaps resistance to ALB evolved rapidly so 1979; Boedo et al., 2012). Among the subspecies of D. carota that no phylogenetic signal could be detected. Tas (2016) in clade A¢ (Fig. 1), subsp. capillifolius and subsp. maximus pres- reported little correlation between ALB disease pheno- ent significantly higher levels of resistance to ALB than the types of carrots and their geographic origins. However, other subspecies, demonstrating new sources of resistance to using the phylogenetic signal proposed here would be ALB for carrot breeding programs. Previous studies showed useful, and stronger levels of association may be obtained. that interspecific crosses between subsp. capillifolius and other Carrot is the second most popular vegetable in the world subspecies have been successful (McCollum, 1975, 1977). A after potato (Heywood, 2014), and it is the economically recent study (Arbizu et al., 2016a) also proposed the use of most valuable member of the Apiaceae. Extensive research subsp. maximus as a new source of genes for the development is being conducted to understand the genetic control of of new carrot cultivars. Further, subsp. carota possess the ALB resistance in carrots (Boiteux et al., 1993; Simon and second highest level of resistance. This result is congruent Strandberg, 1998; Le Clerc and Pawelec, 2009; Le Clerc et with the study conducted by Tas (2016), where 812 acces- al., 2015). Here we provide evidence that taxa belonging to sions of carrots (mainly subsp. carota) were evaluated for ALB clade A (Fig. 2, D. carota subsp. capillifolius and subsp. maxi- disease resistance, concluding that complete resistance was mus) may provide new sources of resistance to ALB. not found, with <40 accessions exhibiting partial resistance to ALB (disease scores £2). Acknowledgments The ability to predict disease resistance to ALB in We are grateful to Kathleen Reitsma and the staff at the North carrots would benefit farmers through reduced use of Central Regional Plant Introduction Station in Ames, IA, for crop science, vol. 57, september–october 2017 www.crops.org 2651 providing germplasm, and Cecile Ané, David Baum, and Lam Cai, X.K., D.M. Spooner, and S.H. Jansky. 2011. A test of taxonomic Ho for discussion and suggestions. This paper represents partial predictivity: Resistance to Potato virus Y in wild relatives of the fulfillment of a Ph.D. degree for CA in Plant Breeding and Plant cultivated potato. Phytopathology 101:1074–1080. doi:10.1094/ Genetics at the University of Wisconsin-Madison. This work PHYTO-02-11-0060 is supported by the USDA-ARS. C. Arbizu is partly funded by Camadro, E.L., M.A. Cauhépé, and P.W. Simon. 2008. Compatibility relations between the edible carrot Daucus carota and D. pusillus, the National Council of Science and Technology of Perú (Con- a related wild species from the Argentinian Pampas. Euphytica cytec, by its initials in Spanish). P. Tas was partly funded by the 159:103–109. doi:10.1007/s10681-007-9462-y USDA-NIFA-Organic Agriculture Research and Extension Chung, Y.S., K. Holmquist, D.M. Spooner, and S.H. Jansky. 2011. Initiative award 2011-51300-30903 (to P.W. Simon). A test of taxonomic and biogeographic predictivity: Resistance to soft rot in wild relatives of cultivated potato. Phytopathology 101:205–212. doi:10.1094/PHYTO-05-10-0139 Conflict of Interest Clotault, J.E., E. Geoffriau, E. Lionneton, M. Briard, and D. Peltier. The authors declare that there is no conflict of interest. 2010. Carotenoid biosynthesis genes provide evidence of geo- graphical subdivision and extensive linkage disequilibrium in the carrot. Theor. Appl. Genet. 121:659–672. doi:10.1007/s00122- Supplemental Material Available 010-1338-1 Supplemental material for this article is available online. de Mendiburu, F. 2016. agricolae: Statistical procedures for agricultural research. R package version 1.2-4. R Foundation for Statistical References Computing. http://CRAN.R-project.org/package=agricolae Arbizu, C., K.R. Reitsma, P.W. Simon, and D.M. Spooner. 2014a. (accessed 22 Oct. 2016). Morphometrics of Daucus (Apiaceae): A counterpart to a phy- du Toit, L.J., F.J. Crowe, M.L. Derie, R.B. Simmons, and G.Q. Pelter. logenomic study. Am. J. Bot. 101:2005–2016. doi:10.3732/ 2005. Bacterial blight in carrot seed crops in the Pacific North- ajb.1400252 west. Plant Dis. 89:896–907. doi:10.1094/PD-89-0896 Arbizu, C., H. Ruess, D. Senalik, P.W. Simon, and D.M. Spooner. Endresen, D.T.F. 2010. Predictive association between trait data 2014b. Phylogenomics of the carrot genus (Daucus, Apiaceae). and ecogeographic data for Nordic barley landrace. Crop Sci. Am. J. Bot. 101:1666–1685. doi:10.3732/ajb.1400106 50:2418–2430. doi:10.2135/cropsci2010.03.0174 Arbizu, C.I., S.L. Ellison, D. Senalik, P.W. Simon, and D.M. Spooner. Endresen, D.T.F., K. Street, M. Mackay, A. Bari, and E. De Pauw. 2016a. Genotyping-by-sequencing provides the discriminating 2011. Predictive association between biotic stress traits and power to investigate the subspecies of Daucus carota (Apiaceae). eco-geographic data for wheat and barley landraces. Crop Sci. BMC Evol. Biol. 16:234. doi:10.1186/s12862-016-0806-x 51:2036–2055. doi:10.2135/cropsci2010.12.0717 Arbizu, C.I., P.W. Simon, F. Martínez-Flores, H. Ruess, M.B. Cre- Ernst, M., C.H. Saslis-Lagoudakis, O.M. Grace, N. Nilsson, H.T. spo, and D.M. Spooner. 2016b. Integrated molecular and mor- Simonsen, J.W. Horn, and N. Rønsted. 2016. Evolutionary pre- phological studies of the Daucus guttatus complex (Apiaceae). Syst. diction of medicinal properties in the genus Euphorbia L. Sci. Bot. 41:479–492. doi:10.1600/036364416X691948 Rep. 6:30531. doi:10.1038/srep30531 Baum, D.A., and S.D. Smith. 2013. Tree thinking: An introduction FAO. 2009. How to feed the world in 2050. FAO. http://www.fao. to phylogenetic biology. Roberts and Co., Greenwood Village, org/f ileadmin/templates/wsfs/docs/expert_paper/How_to_ CO. Feed_the_World_in_2050.pdf (accessed 15 Nov. 2016). Bendiabdellah, A., M. El Amine Dib, N. Meliani, A. Muselli, D. Nas- Farrar, J.J., B.M. Pryor, and R.M. Davis. 2004. Alternaria diseases of sim, B. Tabti, and J. Costa. 2013. Antibacterial activity of Daucus carrot. Plant Dis. 88:776–784. doi:10.1094/PDIS.2004.88.8.776 crinitus essential oils along vegetative life of the plant. J. Chem. Fry, W.E. 1978. Quantification of general resistance of potato cultivars 2013:1–7. doi:10.1155/2013/149502 and fungicide effects for integrated control of potato late blight. Ben-Noon, E., D. Shtienberg, E. Shlevin, H. Vintal, and A. Dinoor. Phytopathology 68:1650–1655. doi:10.1094/Phyto-68-1650 2001. Optimization of chemical suppression of Alternaria dauci, Gawehns, F., B.J.C. Cornelissen, and F.L.W. Takken. 2013. The poten- the causal agent of Alternaria leaf blight in carrots. Plant Dis. tial of effector-target genes in breeding for plant innate immunity. 85:1149–1156. doi:10.1094/PDIS.2001.85.11.1149 Microb. Biotechnol. 6:223–229. doi:10.1111/1751-7915.12023 Boedo, C., S. Benichoua, R. Berruyerc, S. Bersihanda, A. Dongoa, P. Gugino, B.K., J. Carroll, J. Chen, J. Ludwig, and G. Abawi. 2004. Simoneaua et al. 2012. Evaluating aggressiveness and host range Carrot leaf blight diseases and their management in New York. of Alternaria dauci in a controlled environment. Plant Pathol. Cornell Univ. Coop. Ext. http://vegetablemdonline.ppath.cor- 61:63–75. doi:10.1111/j.1365-3059.2011.02494.x nell.edu/factsheets/Carrot_Leaf_Blight.pdf (accessed 25 Nov Boedo, C., R. Berruyer, M. Lecomte, S. Bersihand, M. Briard, V. 2016). Le Clerc et al. 2010. Evaluation of different methods for the Heywood, V.H. 2014. The socio-economic importance of the Apia- characterization of carrot resistance to the alternaria leaf blight les. J. Fac. Pharm. Istanbul Univ. 44:113–130. pathogen (Alternaria dauci) revealed two qualitatively differ- Hijmans, R.J., M. Jacobs, J.B. Bamberg, and D.M. Spooner. 2003. ent resistances. Plant Pathol. 59:368–375. doi:10.1111/j.1365- Frost tolerance in wild potato species: Unraveling the predictiv- 3059.2009.02218.x ity of taxonomic, geographic, and ecological factors. Euphytica Boedo, C., V. Le Clerc, M. Briard, P. Simoneau, M. Chevalier, S. 130:47–59. doi:10.1023/A:1022344327669 Georgeault, and P. Poupard. 2008. Impact of carrot resistance Ho, L.T., and C. Ané. 2014. A linear-time algorithm for Gaussian on development of the Alternaria leaf blight pathogen (Alternaria and non-Gaussian trait evolution models. Syst. Biol. 63:397–408. dauci). Eur. J. Plant Pathol. 121:55–66. doi:10.1007/s10658-007- doi:10.1093/sysbio/syu005 9241-6 Jansky, S., R. Simon, and D.M. Spooner. 2009. A test of taxonomic Boiteux, L.S., P.T. Della Vecchia, and F.J.B. Reifschneider. 1993. Her- predictivity: Resistance to the Colorado potato beetle in wild itability estimate for resistance to Alternaria dauci in carrot. Plant relatives of cultivated potato. J. Econ. Entomol. 102:422–431. Breed. 110:165–167. doi:10.1111/j.1439-0523.1993.tb01230.x doi:10.1603/029.102.0155 2652 www.crops.org crop science, vol. 57, september–october 2017 Jansky, S.H. 2000. Breeding for disease resistance in potato. Plant R Core Team. 2016. R: A language and environment for statistical Breed. Rev. 19:69–155. computing. R Foundation for Statistical Computing, Vienna, Jansky, S.H., R. Simon, and D.M. Spooner. 2006. A test of taxo- Austria. nomic predictivity: Resistance to white mold in wild relatives Revell, L.J., L.J. Harmon, and D.C. Collar. 2008. Phylogenetic of cultivated potato. Crop Sci. 46:2561–2570. doi:10.2135/crop- signal, evolutionary process, and rate. Syst. Biol. 57:591–601. sci2005.12.0461 doi:10.1080/10635150802302427 Jansky, S.H., R. Simon, and D.M. Spooner. 2008. A test of taxo- Rønsted, N., M.R.E. Symonds, T. Birkholm, S.B. Christensen, nomic predictivity: Resistance to early blight in wild relatives A.W. Meerow, M. Molander et al. 2012. Can phylogeny pre- of cultivated potato. Phytopathology 98:680–687. doi:10.1094/ dict chemical diversity and potential medicinal activity of PHYTO-98-6-0680 plants? A case study of amaryllidaceae. BMC Evol. Biol. 12:182. Jeger, M.J., and S.L.H. Viljanen-Rollinson. 2001. The use of the area doi:10.1186/1471-2148-12-182 under the disease-progress curve (AUDPC) to assess quantitative Rubatsky, V.E., C.F. Quiros, and P.W. Simon. 1999. Carrots and disease resistance in crop cultivars. Theor. Appl. Genet. 102:32– related vegetable Umbelliferae. CABI Publishing, New York. 40. doi:10.1007/s001220051615 Sáenz Laín, C. 1981. Research on Daucus L. (Umbelliferae). Anales Jourdan, M., S. Gagné, C. Dubois-Laurent, M. Maghraoui, S. Huet, del Jardín Botánico de Madrid 37:481–533. A. Suel et al. 2015. Carotenoid content and root color of cul- Simon, P.W., R.E. Freeman, J.V. Vieira, L.S. Boiteux, M. Briard, T. tivated carrot: A candidate-gene association study using an Nothnagel et al. 2008. Carrot. In: J. Prohens, M.J. Carena and original broad unstructured population. PLoS One 10:e0116674. F. Nuez, editors, Handbook of crop breeding, Vol. 1: Vegetable doi:10.1371/journal.pone.0116674 breeding. Springer-Verlag, GmBH, Heidelberg, Germany p. Khiutti, A., D.M. Spooner, S.H. Jansky, and D.A. Halterman. 2015. 327–357. Testing taxonomic predictivity of foliar and tuber resistance to Simon, P.W., and J.O. Strandberg. 1998. Diallel analysis of resis- Phytophthora infestans in wild relatives of potato. Phytopathology tance in carrot to Alternaria leaf blight. J. Am. Soc. Hortic. Sci. 105:1198–1205. doi:10.1094/PHYTO-02-15-0046-R 123:412–415. Le Clerc, V., S. Marques, A. Suel, S. Huet, L. Hamama, L. Voisine Soteros, J.J. 1979. Pathogenicity and control of Alternaria radicina and et al. 2015. QTL mapping of carrot resistance to leaf blight with A. dauci in carrots. N. Z. J. Agric. Res. 22:191–196. doi:10.1080/ connected populations: Stability across years and consequences 00288233.1979.10420860 for breeding. Theor. Appl. Genet. 128:2177–2187. doi:10.1007/ Spooner, D.M., M.P. Widrlechner, K.R. Reitsma, D.E. Palmquist, S. s00122-015-2576-z Rouz, Z. Ghrabi-Gammar et al. 2014. Reassessment of practical Le Clerc, V., and A. Pawelec. 2009. Genetic architecture of factors subspecies identifications of the USDA Daucus carota L. germ- underlying partial resistance to alternaria leaf blight in carrot. plasm collection: Morphological data. Crop Sci. 54:706–718. Theor. Appl. Genet. 118:1251–1259. doi:10.1007/s00122-009- doi:10.2135/cropsci2013.04.0231 0978-5 Spooner, D.M., S.H. Jansky, and R. Simon. 2009. Tests of taxonomic Limantseva, L., N. Mironenko, O. Shuvalov, O. Antonova, A. Khiutti, and biogeographic predictivity: Resistance to multiple disease L. Novikova et al. 2014. Characterization of resistance to Glo- and insect pests in wild relatives of cultivated potato. Crop Sci. bodera rostochiensis pathotype Ro1 in cultivated and wild potato 49:1367–1376. doi:10.2135/cropsci2008.04.0211 species accessions from the Vavilov Institute of Plant Industry. Strandberg, J.O. 1983. Infection and colonization of inflorescences Plant Breed. 133:660–665. doi:10.1111/pbr.12195 and mericarps of carrot by Alternaria dauci. Plant Dis. 67:1351– Lovell, D.J., S.R. Parker, T. Hunter, D.J. Royle, and R.R. Coker. 1353. doi:10.1094/PD-67-1351 1997. Influence of crop growth and structure on the risk of Tas, P.M. 2016. Evaluating resistance to Alternaria dauci and related epidemics by Mycosphaerella graminicola (Septoria tritici) in winter traits among diverse germplasm of Daucus carota. Ph.D. diss., wheat. Plant Pathol. 46:126–138. doi:10.1046/j.1365-3059.1997. Univ. of Wisconsin-Madison, Madison, WI. d01-206.x Tavella, C.M. 1978. Date of heading and plant height of wheat variet- McCollum, G.D. 1975. Interspecific hybrid Daucus carota ´ D. capil- ies, as related to septoria leaf blotch damage. Euphytica 27:577– lifolius. Bot. Gaz. 136:201–206. doi:10.1086/336803 580. doi:10.1007/BF00043184 McCollum, G.D. 1977. Hybrids of Daucus gingidium with cultivated Turner, S.D., P.L. Maurizio, W. Valdar, B.S. Yandell, and P.W. Simon. carrots (D. carota subsp. sativus) and D. capillifolius. Bot. Gaz. 2017. Dissecting the genetic architecture of shoot growth in 138:56–63. doi:10.1086/336899 carrot (Daucus carota L.) using a diallel mating design. bioRxiv Miedaner, T., and V. Korzun. 2012. Marker-assisted selection for dis- 115519 doi:10.1101/115519 ease resistance in wheat and barley breeding. Phytopathology Turner, S.D., N.D. Miller, E.P. Spalding, and P.W. Simon. 2016. 102:560–566. doi:10.1094/PHYTO-05-11-0157 Image-based phenotyping and Bayesian analysis of a diallel mat- Mourtzinis, S., F.J. Arriaga, K.S. Balkcom, and B.V. Ortiz. 2013. Corn ing design for top size in carrot (Daucus carota L.). Nat. Assoc. grain and stover yield prediction at R1 growth stage. Agron. J. Plant Breeders. https://www.plantbreeding.org/files/napb/ 105:1045–1050. doi:10.2134/agronj2012.0393 annual-meeting-2016-abstract-book.pdf (accessed 25 Nov 2016). Neergaard, P. 1977. Seed pathology. Halsted Press, NewYork. Vintal, H., E. Ben-Noon, E. Shlevin, U. Yermiyahu, D. Shtienberg, doi:10.1007/978-1-349-02842-9 and A. Dinoor. 1999. Influence of rate of soil fertilization on Peeters, J.P., H.G. Wilkes, and N.W. Galwey. 1990. The use of eco- alternaria leaf blight (Alternaria dauci) in carrots. Phytoparasitica geographic data in the exploitation of variation from gene banks. 27:193–200. doi:10.1007/BF02981458 Theor. Appl. Genet. 80:110–112. doi:10.1007/BF00224023 Wickman, H. 2011. The split-apply-combine strategy for data analy- Piquerez, S.J.M., S.E. Harvey, J.L. Beynon, and V. Ntoukakis. 2014. sis. J. Stat. Softw. 40:1–29. Improving crop disease resistance: Lessons from research on Arabidopsis and tomato. Front. Plant Sci. 5:671. doi:10.3389/ fpls.2014.00671 crop science, vol. 57, september–october 2017 www.crops.org 2653