In the context of this work, Cluster A.1 corresponds to a well-supported clade (100%, 1.00; Fig. 4) and comprises 36 sequences of which 23 are ASVs with 190 entries (Suppl. material 1: Table S2), while the expanded/detailed analysis includes a total of 297 entries (Suppl. material 3: Fig. S2a). Cluster A.1 is formed by material deriving from temperate regions of the Northern Hemisphere (except of two sequences originating from Argentinian specimens) on a large range of host plants under various names (Suppl. material 1: Table S2; Fig. 4). Interspecific genetic distances within Cluster A.1 are very low (i.e. 0.008 ± 0.004) in comparison to values calculated in other Ganoderma groups, which, in conjunction with the high sequence similarity values noted (98.96 ± 0.62%), are indicative of low divergence amongst taxa. Therefore, the criteria set in this study for phylogenetic species are not generally met for members of Cluster A.1 since a significant overlap exists between intraspecific and interspecific variability (absence of a barcoding gap; Fig. 8). Hence, in this particular case, ITS demonstrates poor species-level resolution and delimitation of taxa seems to be supported by multigene approaches only (Zhou et al. 2015; Loyd et al. 2018; Ye et al. 2019).

The major group of Cluster A.1 corresponds to G. lucidum (Curtis) P. Karst. (G. lucidum sensu stricto), which is represented by 10 sequences (or 97 entries) in the main dataset (Fig. 4). The majority of sequences which grouped within this species were found to be accurately identified (107 out of the 153 entries; Table 1); the highest number of erroneously-labelled sequences placed in G. lucidum were under the name “G. tsugae” (12 in total, six of which derived from China). The majority of G. lucidum sequences derive from material originating from Eurasia, growing mainly on hardwoods with some occasional reports of occurrence on conifers, for example, Larix and Pinus spp. (Suppl. material 1: Table S2). An exception was formed by nine entries from the US (California and Utah); they most probably correspond to introduced material (Loyd et al. 2018), as well as two sequences from Argentina under the name “G. oerstedii” (Moncalvo et al. 1995a). On the basis of the outcome of the present study, the latter were apparently misidentified and the respective material belongs to G. lucidum; its existence in Argentina could probably be attributed to human-mediated transfer. A distinct subclade within G. lucidum consists of 18 entries corresponding to three ASVs and three singletons (75%, 0.97; Fig. 4) from East Asian collections only (Park et al. 2012; Zhou et al. 2015; Xing et al. 2018); two conserved substitutions in both ITS spacers differentiate the respective sequences from the rest of G. lucidum (position 225 in ITS1 and position 378 in ITS2; Table 3; Suppl. material 2: Fig. S1).

G. carnosum Pat. was first described on Abies in southwest France (Pyrenees Mts.) and is distributed throughout Europe (Jahn et al. 1980, 1986; Mattock 2001; Ryvarden and Melo 2017); no reports exist of its occurrence in other continents. With respect to morphology, it resembles G. lucidum and the most prominent discriminating characters are the preference for conifers (Abies and Pinus spp.) and the blackish shiny upper surface when mature. The wider and highly-granulose (rougher) basidiospores, as well as the pore density, are also referred to as being of diagnostic value in taxonomic keys (Kotlaba and Pouzar 1993; Ryvarden and Gilbertson 1993; Ryvarden and Melo 2017). As a result of the ITS meta-analysis, 26 entries of G. carnosum originating from Europe (five generated in this study; Table 1 and Fig. 4) formed a group demonstrating high intraspecific sequence similarity. Most of them were accurately labelled (22 out of 32 G. carnosum entries available in public databases), while four were initially determined as “G. lucidum” (Table 1). The phylogenetically closest taxa to G. carnosum are G. oregonense and G. lucidum with which high values of sequence similarity were noted (99.68 ± 0.20% and 99.49 ± 0.24%, respectively).

Sequences deriving from two U.K. specimens (deposited as G. carnosum and G. lucidum) and from two environmental samples (Estonia, labelled as “uncultured soil fungus”) formed a distinct group (93%, 1.00; Fig. 4). It is provisionally named “G. aff. carnosum” due to the initial sequence labelling and European distribution and presents high sequence affinity with G. oregonense (98.87 ± 0.22%) and G. carnosum (98.84 ± 0.16%).

G. oregonense Murrill and G. tsugae Murrill are closely-associated taxa (Adaskaveg and Gilbertson 1986, 1988; Moncalvo 2000). They are mainly distinguished on the basis of basidiospores size and geographic distribution; the former is mainly found in western USA in temperate woods dominated by Tsuga heterophylla, Pseudotsuga menziesii, Picea spp. and Abies spp. However, the inclusion of environmental samples in this study expanded the known distribution of this taxon to Europe (Estonia, UDB0287378). G. oregonense is represented by four sequences corresponding to 21 entries (Fig. 4) in our analysis, while six singletons are also associated with this taxon and are included in the expanded dataset (Suppl. material 1: Table S2 and Suppl. material 3: Fig. S2a). On the other hand, G. tsugae is mainly recorded in eastern USA in temperate hemlock forests (Loyd et al. 2018), while it is also reported to occur in Canada (Adaskaveg and Gilbertson 1986, Ryvarden and Gilbertson 1993) and in Asia (Zhao 1989; Pegler and Yao 1996; Zhao and Zang 2000). Such reports were confirmed through the outcome of the present work (Table 1). G. tsugae is hereby represented by nine sequences corresponding to 41 entries in the databases examined (Fig. 4), while another 13 singletons were added in the expanded analysis (Suppl. material 1: Table S2 and Suppl. material 3: Fig. S2a). Most of them were found to be correctly identified (46 entries out of 54; Table 1; a sequence from the type specimen was also included, Z37054/Z37079). The ITS sequence similarity and genetic distance between G. tsugae and G. oregonense was found to be high and low, respectively (99.19 ± 0.62% and 0.005 ± 0.003, respectively), demonstrating that their distinct phylogenetic status cannot be evidenced through the use of ITS alone. In addition, the boxplot analysis revealed the absence of a barcoding gap between the two taxa (Fig. 8). However, application of a multigene approach permitted their delimitation (Loyd et al. 2018).

In the past, G. tsugae was occasionally reported to be conspecific with G. carnosum or G. lucidum on the basis of morphological observations (Donk 1974; Ryvarden and Melo 2017). The analysis performed in this work separates G. tsugae from G. carnosum and G. lucidum, which is in agreement with recent reports (Zhou et al. 2015; Loyd et al. 2018), although the interspecific sequence similarity values are high, i.e. 99.15 ± 0.58% and 98.91 ± 0.61%, respectively. In addition, a unique sequence, initially labelled “G. carnosum” (Z37057/Z37082; strain JAHN 1197-121, Germany), represents an authentic specimen of G. atkinsonii Jahn, Kotl. & Pouzar, which was later synonymised with G. carnosum (Jahn et al. 1986) and maintained as such by Moncalvo et al. (1995). The outcome of the present study shows that it forms part of G. tsugae (Suppl. material 1: Table S2 and Suppl. material 3: Fig. S2a).

Noteworthy cases pertain to sequences under the names of “G. valesiacum” and “G. ahmadii” which grouped within G. tsugae (Table 1; Fig. 4). G. valesiacum Boud. was described by Boudier and Fischer (1895) from a collection on Larix in Valais, Switzerland. The type specimen is almost destroyed; however, remains of the cutis are similar to the hymenodermiform anatomy presented by G. lucidum, but with smaller basidiospores possessing pronounced surface cambers and a “white, punk-like context” under a sometimes cracking-crust (Ryvarden and Gilbertson 1993; Ryvarden and Melo 2017). Three ITS sequences, labelled as “G. valesiacum”, are found in GenBank and were included in this work: (a) Z37056/Z37081 (strain CBS 282.33; UK), also used in previous studies (Moncalvo et al. 1995a; Hong and Yung 2004), is grouped within G. tsugae and presents sequence similarity values higher than 97.63% to other members of this taxon; (b) MG711807 (strain LE-BIN 2350; Russia, Altai Mts.) forms a clade within G. lucidum presenting an identical sequence with the most common ASV of this species (i.e. MG706224); (c) JQ520218 is grouped within G. sessile possessing the same sequence as the most common ASV of this species (i.e. MG654214; Suppl. material 1: Table S2); this particular sequence is erroneously associated with strain CBS 428.84 (USA), which had been earlier sequenced and correctly labelled as G. tsugae (X78735/X78756; Moncalvo et al. 1995b). Hence, molecular evidence from sequences labelled as “G. valesiacum” is in accordance with the views expressed in previous studies stating that such material is conspecific with either G. lucidum or G. tsugae on the basis of morphological features alone (Steyaert 1972; Stalpers 1978; Adaskaveg and Gilbertson 1986; Nunez and Ryvarden 2000; Wasser et al. 2006). Still, in the absence of a sequence from the holotype, it is not possible to draw definite conclusions on whether this name corresponds to a valid phylogenetic species. On the other hand, the only ITS sequence representing G. ahmadii Steyaert derives from the type material (strain FWP14329, Pakistan; Z37047/Z37098) and presents relatively-high sequence similarity values to members of the G. tsugae group (98.63 ± 0.36%). Therefore, on the basis of data available, the status of G. ahmadii remains ambiguous and additional specimens need to be studied.

In the frame of this study, four ASVs representing 20 entries formed a well-supported group (81%, 1.00; Fig. 4) including type material of two Chinese taxa recently described, i.e. G. leucocontextum T.H. Li, W.Q. Deng, Sheng H. Wu, D.M. Wang & H.P. Hu and G. weixiensis Ye et al. (Li et al. 2014; Ye et al. 2019). The two taxa presented high intraspecific sequence similarity indicative of their close geographic origin (i.e. 99.64 ± 0.28%); all studied specimens originated from China, Tibet and Nepal. Hence, ITS data alone could not discriminate the two taxa (Fig. 4) and the use of additional markers was necessary for establishing the latter species (Ye et al. 2019).

Cluster A.2 (94%, 1.00; Fig. 4) is represented by 73 unique sequences; 47 are ASVs deriving from 309 entries in the databases (Table 1 and Suppl. material 1: Table S2). The expanded/detailed tree is formed by 263 sequences representing 517 entries in the databases (Suppl. material 1: Table S2 and Suppl. material 4: Fig. S2b).

A major subclade is formed by specimens collected in southeast Asia and Australia growing mostly on angiosperms (93%, 1.00 and 72%, 1.00, Fig. 4 and Suppl. material 4: Fig. S2b, respectively). The sequences had been deposited under different taxonomic names, i.e. G. weberianum (Bres. & Henn. ex Sacc.) Steyaert, G. sichuanense J.D. Zhao & X.Q. Zhang, G. tenue J.D. Zhao, L.W. Hsu & X.Q. Zhang, G. lucidum, G. microsporum R.S. Hseu and Ganoderma sp. and present rather low heterogeneity as evidenced by the values of sequence similarity and genetic distance obtained (98.87 ± 0.62% and 0.010 ± 0.005, respectively). We consider that this particular subclade is related to G. weberianum sensu Steyaert (Moncalvo 2000) and includes at least two not adequately (by ITS alone) resolved groups. One of them corresponds to G. sichuanense and consists of 19 entries originally identified as G. sichuanense (9), G. tenue (2), “G. weberianum” (4), “G. lucidum” (2), Ganoderma sp. (1) and “uncultured soil fungus” (1), which derive from specimens collected in southeast Asia and Australia on angiosperms and gymnosperms (Table 1, Suppl. material 1: Table S2). This taxon demonstrates high sequence similarity (99.40 ± 0.22%) and low genetic distance values (0.006 ± 0.003). Two sequences assigned to G. sichuanense (including the holotype: HMAS 42798, JQ781877; Cao et al. 2012) group with sequences identified as G. tenue (Figure 3 and Suppl. material 4: Figure S2b); unfortunately, no sequence is available from the type material of G. tenue (Zhao et al. 1979a), which makes it difficult to comment on its relationships to associated taxa. The other group is composed of 12 sequences [originally identified as G. weberianum (9) and G. microsporum (corresponding to the type, RSH 0821; Moncalvo et al. 1995b)] which are placed together (91%, 0.99; Fig. 4) and demonstrate high intraspecific sequence similarity (99.16 ± 0.48%). On the basis of the aforementioned findings, we consider that they form part of G. weberianum sensu stricto (originally described from Samoa; Steyaert 1972). Moreover, although it is not possible to draw any definite conclusions about the status of G. microsporum, we tend to agree with the view expressed by Smith and Sivasithamparam (2000) that it is a subspecific entity within G. weberianum. The two groups corresponding to G. sichuanense and G. weberianum demonstrate interspecific ITS sequence similarity and genetic distance values (Fig. 8) which do not support the existence of two distinct phylospecies on the basis of the criteria hereby set.

G. hoehnelianum Bres. forms a monophyletic group in Cluster A.2 (99%, 1.00; Table 1, Fig. 4) and includes 12 entries deriving from southeast Asian (China, Myanmar) and African (Gabon) material; sequences of the two origins are separated into two subgroups (Fig. 4), but they exhibit low variability (sequence similarity: 99.63 ± 0.23%). On the other hand, G. austroafricanum Coetzee, M.J. Wingf., Marinc., Blanchette is also a well-supported species (87%, 0.99; Suppl. material 4: Fig. S2b) represented by two singletons, including the type specimen from South Africa. Moreover, G. carocalcareum Douanla-Meli forms a subclade (67%, 0.99; Suppl. material 4: Fig. S2b) consisting of 11 sequences which originate from material collected in Africa (Table 1); two are under this name (including the type specimen), while the rest are labelled either as “G. weberianum” (3) or as Ganoderma sp. (1) and form a well-supported subgroup within the clade (88%, 0.99; Fig. 4). However, sequence similarity and genetic distance values (98.61 ± 0.36% and 0.013 ± 0.002, respectively) do not permit their dinctinction. The entry MK603806 (MUCL 49272) represents the “C2.2” clade in the study of Cabarroi-Hernández et al. (2019) which is composed of several MUCL specimens from Cameroon and Gabon. Genetic distance and sequence similarity values support the distinct status of G. carocalcareum from G. weberianum (0.022 ± 0.006 and 97.50±0.78%, respectively).

A new phylogenetic species within Cluster A.2 is hereby proposed and is provisionally named “Ganoderma sp. A1” (corresponding to the UNITE DOIs SH1740420.08FU, SH1740444.08FU and SH1740445.08FU); its monophyly is strongly supported in both trees (100%, 1.00; Fig. 4 and Suppl. material 4: Fig. S2b). It is represented by eight unique sequences (i.e. four ASVs and four singletons; Suppl. material 1: Table S2 and Suppl. material 4: Fig. S2b) corresponding to 17 entries. On the basis of available information, all sequences derive from specimens initially identified as “G. weberianum” originating from India on a large range of eudicots (Suppl. material 1: Table S2). Ganoderma sp. A1 presents relatively-low intraspecific genetic distances (0.010 ± 0.003) and high sequence similarity (99.01 ± 0.18%).

Other four entries of dubious identity derive from material originating from Asia (China and India) labelled as Ganoderma sp., “G. weberianum” (2) and “Ganoderma cf. weberianum”, as well as from Brazil under the name “G. subamboinense” (Table 1 and Suppl. material 1: Table S2) and exhibit low genetic distance and high sequence similarity (0.006 ± 0.003 and 99.18 ± 0.43%, respectively). They form part of the same larger subclade together with Ganoderma sp. A1, G. mexicanum and G. parvulum (Fig. 4); however, they do not retain the same position in the expanded tree (Suppl. material 4: Fig. S2b). For the purposes of this work, we provisionally maintain them as Ganoderma aff. weberianum (corresponding to the UNITE DOI SH1723064.08FU) since it is not possible to determine their exact identity from the data available.

Another major subclade consisting of material originating from the Neotropics is formed by a total of 17 entries. Four of them are singletons, while the other 13 are grouped in six ASVs (Suppl. material 1: Table S2; Fig. 4); the expanded/detailed tree includes 38 entries corresponding to 31 sequences (Suppl. material 4: Fig. S2b). They were deposited under several names: G. parvulum Murrill (10 sequences), G. subamboinense var. laevisporum Bazzalo & J.E. Wright (7, including type material), G. mexicanum Pat. (6), G. subamboinense (Henn.) Bazzalo & J.E. Wright ex Moncalvo and Ryvarden (3, including type material), “G. tuberculosum” (1), “G. weberianum” (5), G. sessiliforme Murrill (1), G. stipitatum (Murrill) Murrill (1) and Ganoderma sp. (4). The outcome of this work shows that this material forms part of two terminal groups corresponding to G. mexicanum and G. parvulum (Fig. 4) in accordance with the findings of a recent study by Cabarroi-Hernández et al. (2019). Hence, the former is selected as the earliest valid name to accommodate specimens also reported as “G. sessiliforme” and “G. subamboinense var. laevisporum” from Argentina, Brazil, Martinique, Mexico and USA, while the latter for material also labelled as “G. subamboinense”, “G. subamboinense var. subamboinense” and “G. stipitatum” from Brazil, Colombia, Costa Rica, Cuba, French Guiana and USA (77%; Fig. 4). Such a separation is in accordance with morphological observations made by Ryvarden (2000; 2004) and Cabarroi-Hernández et al. (2019). However, the support that this group of taxa receives by ITS data is not adequate (Suppl. material 4: Fig. S2b) and the respective interspecific sequence similarity and genetic distance values (Fig. 8) do not separate them on the basis of the criteria hereby set.

Another well-supported terminal clade (100%, 1.00; Fig. 4, Suppl. material 4: Fig. S2b) is formed by two entries labelled as “G. resinaceum” and Ganoderma sp. deriving from China. It is provisionally named Ganoderma sp. A2 (corresponding to the UNITE DOI SH2762559.08FU) since it complies with the criteria set in this study and it is hence considered as a new phylospecies. The genetic distance and sequence similarity values versus the most closely-related species (i.e. G. resinaceum) are 0.028 ± 0.003 and 97.52 ± 0.23%, respectively.

G. resinaceum Boud. is represented by 10 ASVs and 65 entries (Fig. 4), while the expanded/detailed tree includes 65 sequences corresponding to 126 entries (Suppl. material 4: Fig. S2b). Intraspecific genetic distance and sequence similarity values are well within the respective ranges observed for taxa of Clade A (i.e. 0.005 ± 0.003 and 99.46 ± 0.31%, respectively). This species appears to be very common throughout Europe (type locality), but it also occurs in Asia (e.g. China, India, Iran, Iraq, South Korea and Turkey) and Africa (Egypt, South Africa and Tunisia), being reported on a wide range of angiosperms (Suppl. material 1: Table S2). On the other hand, there is great controversy regarding the existence of G. resinaceum in the Americas. Its occurrence was reported by several authors in the past, for example, in Mexico (Torres-Torres et al. 2015), Brazil (Loguercio-Leite et al. 2005) and Argentina (Bazzalo and Wright 1982), although it was admittedly confused with other species names (e.g. G. oerstedii, G. parvulum or G. subincrustatum; de Lima Júnior et al. 2014; Torres-Torres et al. 2015). In addition, it has been synonymised with Neotropical taxa, such as G. chaffangeonii Pat. (type locality Venezuela; Steyaert 1980; Torres-Torres et al. 2012) and G. praelongum Murrill (type locality Cuba). However, no report of its alleged existence in America is so far supported by DNA data. Moreover, in the frame of this study, none of the sequences examined and confirmed to be G. resinaceum represents material originating from this continent. Therefore and to the best of our knowledge, the presence of G. resinaceum in the Americas cannot be confirmed by molecular evidence and its distribution seems to be restricted to the Old World.

Another new phylogenetic species is hereby proposed, provisionally named “Ganoderma sp. A3” (corresponding to the UNITE DOI SH1723084.08FU). It is strongly supported in both trees (100%, 1.00 and 99%, 1.00 in Fig. 4 and Suppl. material 4: Fig. S2b, respectively) and is represented by twelve singletons under the names “G. resinaceum” (5), “Ganoderma cf. resinaceum” (3) Ganoderma sp. (2), “G. lucidum” (1) and “uncultured Ganoderma” (1), all deriving from material from east Asia (Table 1 and Suppl. material 1: Table S2). This distinct taxon (intraspecific genetic distance: 0.007 ± 0.002; sequence similarity: 99.22 ± 0.37%) presents at least six different conserved positions in ITS1 and in ITS2 when compared to G. resinaceum (Suppl. material 3: Fig. S1), which is the closest phylogenetic relative (genetic distance: 0.024 ± 0.005; sequence similarity: 97.59 ± 0.51%).

G. sessile Murrill is a well-supported (78%, 1.00; Fig. 4) and highly-represented species in terms of deposited sequences (227; Suppl. material 1: Table S2) showing high intraspecific sequence similarity (99.73 ± 0.14%). Many of the sequences were initially labelled as “G. resinaceum” (60 entries) and “G. lucidum” (9). Most G. sessile sequences derived from specimens collected in the USA, while the rest originate from Asia (including the Caucasus area) and one from Argentina (JQ520199; originally labelled as G. resinaceum), possibly as an outcome of human-mediated transfer.

A closely-related and well-supported (93%, 1.00; Fig. 4) group is formed by sequences belonging to G. polychromum (Copel.) Murrill. This species is represented by 19 entries deriving from material collected in either western USA (California, Washington and Oregon; associated with hardwood) or India; some of them were deposited under the names “G. lucidum” (6) or “G. sessile” (2) (Table 1). However, pairwise comparisons of G. polychromum sequences with those belonging to closely-positioned taxa revealed rather low genetic distance and high sequence similarity values (for example, vs. G. sessile: 0.012 ± 0.004, 98.76 ± 0.32%, respectively).

A distinct phylogenetic group (98%, 1.00; Fig. 4), sister to G. polychromum, consists of seven sequences presenting high intraspecific similarity, i.e. 98.88 ± 0.64%; the respective material derived the Americas and it appears under the names “G. platense”, “G. resinaceum”, “G. sessile” “G. cf. sessile” and “G. zonatum” (Table 1). Amongst those, the only one that could possibly have been correctly identified corresponds to G. platense Speg. because the correct topologies of G. resinaceum and G. sessile are found elsewhere within Cluster A.2, while G. zonatum forms part of Clade D (Fig. 4). However, since very limited information is available concerning this particular specimen (G. platense isolate BAFC384, AH008109; Gottlieb et al. 2000), we prefer not to draw any conclusions concerning the name of this terminal subclade. In addition, because of the close affinity it presents with G. polychromum (genetic distance: 0.014 ± 0.004, sequence similarity: 98.10 ± 0.42%), it does not abide (albeit marginally) with the criteria we set in order to be characterised as a distinct phylogenetic species; therefore, it is provisionally named “G. aff. polychromum” (corresponding to the UNITE DOIs SH1723162.08FU and SH1723226.08FU).

Another terminal group, although not adequately supported in any of the trees constructed, forms a sister clade to G. sessile/G. polychromum (Fig. 4, Suppl. material 4: Fig. S2b) and includes four sequences originally identified as “G. lucidum” deriving from India (3) and Turkey on eudicots (i.e. Cassia fistula and Tamarindus indica). It presents low genetic distance and high sequence similarity values in pairwise comparisons to G. sessile (0.014 ± 0.002 and 97.34 ± 0.16%); therefore, its status is dubious and it is provisionally named “G. aff. sessile” (corresponding to the UNITE DOI SH1723202.08FU).

Cluster A.3 (76%, 1.00; Fig. 5) includes material originating from south and east Asia, tropical Africa, Australia and America (no occurrence in Europe) growing mostly on eudicot hosts (Suppl. material 5: Fig. S2c). It is composed of a total of 1385 entries, 889 of which are grouped in 145 ASVs (Suppl. material 1: Table S2) and comprises 19 terminal groups, 14 of which are well-supported in the present analysis.

G. tuberculosum Murrill is strongly supported in the generated trees (100%, 1.00; Fig. 5 and Suppl. material 5: Fig. S2c) and is represented by 36 sequences, 23 of which are grouped in six ASVs (Suppl. material 1: Table S2 and Suppl. material 5: Fig. S2c). Most of the respective material was correctly labelled (28 entries) and collections derive entirely from the Neotropics, i.e. southeast USA, Cuba, Martinique, Panama, Colombia and Brazil (Table 1). Murrill’s type material of G. tuberculosum from British Honduras was very similar to basidiomes from Martinique examined by Welti and Courtecuisse (2010), thus confirming the wider distribution of this species in the Caribbean as indicated by our analysis. It is worth noting that sequences from Colombia (2) and Texas (1) form a distinct strongly-supported subgroup (99%, 1.00; Suppl. material 5: Fig. S2c) presenting conserved differences in ITS sequences (three to five positions in ITS1 and three in ITS2; Table 3 and Suppl. material 3: Fig. S1); however, their similarity values to the other G. tuberculosum sequences are high (i.e. 99.27 ± 0.58%) and do not seem to support a distinct species status. On another issue, recent studies reported the presence of G. oerstedii (Fr.) Murrill in Mexico, Costa Rica and Honduras (amongst other areas in the Neotropics) and referred to diagnostic characters, such as “the color of the basidiomata, context with resinous bands, cuticle cells with protuberances and/or branches and partially anastomosed basidiospore pillars” (Mendoza et al. 2011; Torres-Torres et al. 2015). Both Ryvarden (2000) and Torres-Torres et al. (2015) considered G. tuberculosum as a synonym of G. oerstedii, but this view cannot be supported (but neither contradicted) by the findings of the present study, since sequence data from the holotype or of a correctly-designated epitype of G. oerstedii are missing.

Furthermore, one sequence (JX310812) labelled as “G. chalceum” originating from Brazilian material and one sequence under the name G. concinnum Ryvarden (possibly of South American origin) form a terminal subclade which nested close to G. tuberculosum (0.97; Fig. 5). The two sequences present relatively-low sequence similarity (98.01%) and rather high genetic distance (0.020) and their conspecificity is uncertain. Since G. chalceum (Cooke) Steyaert is a species described on the basis of material originating from Africa (type specimen from Sierra Leone), southeast Asia and Oceania (Steyaert 1967) and because the only other ITS sequence available under this name is grouped in Clade D of the present study (together with other entries originating from Africa), the real identity of JX310812 remains ambiguous. Therefore, we prefer to use the name G. concinnum for describing this particular group.

Two sister subclades (0.99, Suppl. material 5: Fig. S2c) correspond to G. flexipes Pat. and G. wiiroense E.C. Otto, Blanchette, C.W. Barnes & Held. Both are strongly supported in the generated trees (1.00%, 1.00; Fig. 5 and Suppl. material 5: Fig. S2c). The former consists of seven sequences deriving from material collected on both angiosperms and gymnosperms in tropical Asia (China, Laos and Vietnam; Suppl. material 1: Table S2), presenting relatively-high intraspecies genetic distance (0.011 ± 0.005) which is in agreement with the formation of two well-supported terminal subgroups. The latter also constitutes a monophyletic species represented by 14 entries originating from Ghana (including the type specimen), Senegal and India.

Two new well-supported monophyletic species, provisionally named as “Ganoderma sp. A4” (no UNITE DOI available) and “Ganoderma sp. A5” (corresponding to the UNITE DOI SH1723120.08FU) are revealed in this study. Ganoderma sp. A4 is represented by two sequences deriving from Argentinian material which were originally identified as “G. lucidum” (100%, 1.00; Fig. 5 and Suppl. material 5: Fig. S2c). Ganoderma sp. A5 comprises seven sequences labelled as “G. multiplicatum”, all originating from Asian specimens (100%, 1.00; Fig. 5 and Suppl. material 5: Fig. S2c); this name is apparently misapplied since G. multiplicatum occurs in the Neotropics and corresponds to another, rather distant, terminal group. Both Ganoderma sp. A3 and Ganoderma sp. A4 present high intraspecific sequence similarities (> 98.90%), whereas their respective interspecific values are indicative of their distinct status (93.64 ± 0.29% and 0.058 ± 0.03).

G. philippii (Bres. & Henn. ex Sacc.) Bres. constitutes a well-supported species (97%, 1.00; Fig. 5) represented by 102 sequences, of which 69 are grouped in 13 ASVs (Suppl. material 1: Table S2 and Suppl. material 5: Fig. S2c). The respective material originates from south and southeast Asia (China, Indonesia, Malaysia and Thailand) and is mainly deposited under the synonym G. pseudoferreum (Wakef.) Overeem & B.A. Steinm. Intraspecific genetic distance and sequence similarity values lie well within the ranges observed for species of Cluster A (i.e. 0.005 ± 0.003, max. 0.009; 99.44 ± 0.25%, min. 98.90%).

Three species, i.e. G. lingzhi S.H. Wu, Y. Cao & Y.C. Dai, G. ravenelii Steyaert and G. curtisii (Berk.) Murrill, form a strongly-supported group (A.3.1; 98%, 1.00; Fig. 5). Amongst them, G. lingzhi is distributed in south and east Asia on a wide range of angiosperms (Suppl. material 1: Table S2) and corresponds to a well-supported subclade (100%, 1.00; Fig. 5). It is represented by the largest number of entries for any given species in the present study (615, ca. 16% of the total generic entries; Table 1); 436 of them are grouped in 41 ASVs, while the dominant ASV is represented by 105 identical ITS entries (Suppl. material 1: Tables S2 and S4 and Suppl. material 5: Fig. S2c). Only 54% of these entries were deposited as G. lingzhi (including the holotype, JQ781858), while many of them were originally labelled as “G. lucidum” (206) since collections of the traditional medicinal fungus ‘Lingzhi’ were mainly identified as such for many years. Cao et al. (2012) finally proposed the name G. lingzhi for this fungus and thereby marked the onset of a debate concerning the real identity (correct name) of the fungus, i.e. G. sichuanense or G. lingzhi. Wang et al. (2012) and Yao et al. (2013) supported the former view by sequencing a so-called epitype (voucher HMAS252081; KC662402) of the original material. However, the epitypification of G. sichuanense did not comply with the International Code of Nomenclature for algae, fungi and plants (Zhou et al. 2015). Hence, this particular collection of “G. sichuanense” corresponds to G. lingzhi as was recently explained (Dai et al. 2017). On the other hand, the holotype represents the true G. sichuanense which is not related to G. lingzhi (Cao et al. 2012; Liao et al. 2015; this work) and is positioned in Cluster A.2. It is of interest that, although a large number of G. lingzhi sequences were included in this meta-analysis, intraspecific genetic divergence was low (0.004 ± 0.003, 99.42 ± 0.32%).

G. curtisii (70%, 1.00; Suppl. material 5: Fig. S2c) is closely related to G. lingzhi in agreement with previous phylogenies (Zhou et al. 2015; Thawthong et al. 2017). It occurs in North America – being widespread in the eastern parts of the USA – primarily on angiosperms. In the context of the present study, 142 entries were grouped within this species; 89 of them correspond to 20 ASVs (Suppl. material 1: Table S2 and Suppl. material 5: Fig. S2c). The majority (124) were deposited with the correct name in databases, while nine were labelled as G. meredithae Adask. & Gilb. (including the type material; MH862131) and three as G. curtisii f. sp. meredithae. However, G. meredithae is considered to be a synonym of G. curtisii (Moncalvo 2000) as was later confirmed (Loyd et al. 2018). G. curtisii demonstrates low levels of variability evidenced by its intraspecific genetic distance and sequence similarity values (0.003 ± 0.001 and 99.49 ± 0.22%, respectively). On the other hand, G. ravenelii presents an overlapping geographic distribution with G. curtisii, for example, in south and east USA (Florida and North Carolina; Suppl. material 1: Table S2) and it forms a well-supported terminal subgroup (78%, 1.00; Fig. 5) being composed of 12 sequences from specimens isolated on both angiosperms and gymnosperms (Suppl. material 5: Fig. S2c). However, the outcome of the present work shows that these two taxa exhibit high affinity (genetic distance: 0.011 ± 0.003; sequence similarity: 98.93 ± 0.27%) and no clear ITS barcoding gap is evident between them (Fig. 8); their delineation is adequately supported only through the application of a multigene approach (Loyd et al. 2018).

Six other closely-related species are found in Cluster A.3 (i.e. G. multiplicatum, G. destructans, G. steyaertanum, G. mizoramense, G. martinicense and G. multipileum) (Fig. 5, Suppl. material 5: Fig. S2c). Amongst them, G. multiplicatum (Mont.) Pat. was originally described from French Guiana (Moncalvo and Ryvarden 1997) and its presence was evidenced in several other areas of South America (Ryvarden 2004; de Lima Júnior et al. 2014; Bolaños et al. 2016; Torres-Torres et al. 2012). Pertinent material, analysed in this study, formed an external subclade with strong support (99%, 1.00; Fig. 5); genetic distance and sequence similarity values (0.002 ± 0.001 and 99.54 ± 0.25%, respectively) are indicative of low intraspecific variability. All seventeen sequences derived from specimens collected in the Neotropics (Mexico, Brazil and Colombia; Table 1); seven of them were initially labelled as “G. perzonatum”. However, the exact status of G. perzonatum Murrill remains ambiguous; it was originally described from Cuba and its morphological features associate it with G. parvulum (Cluster A.2 in this work) as reported by Moncalvo and Ryvarden (1997) and Cabarroi-Hernández et al. (2019). According to the latter study, “G. perzonatum could represent another closely related taxon in the vicinity of G. mexicanum / G. parvulum”; hence, examination of additional specimens is needed to arrive at robust conclusions. G. multiplicatum was also recorded in Asia and Africa (Steyaert 1980; Zhao 1989; Wang and Wu 2007; Bhosle et al. 2010). To the best of our knowledge, the occurrence of this species in a region other than the Neotropics was never evidenced through the use of molecular data; the four sequences from China under this name correspond to a distinct phylospecies nested in Cluster A.3, i.e. Ganoderma sp. A5 (Table 1, Fig. 5) as previously explained. Therefore, the Asian material that groups in Ganoderma sp. A5 does not seem to be associated with the name G. multiplicatum in contrast to what was recently reported (Hapuarachchi et al. 2019).

In this study, G. destructans M.P.A. Coetzee, Marinc., M.J. Wingf. represented by 39 entries (including the type material; Table 1) is grouped together with G. dunense Tchotet, Rajchenb. & Jol. Roux; the latter consists of three entries (one from the type material) which are identical to G. destructans sequences (Table 1, Fig. 5 and Suppl. material 5: Fig. S2c). Specimens of both taxa originate from South Africa on eudicots (Coetzee et al. 2015, Tchotet Tchoumi et al. 2018). ITS sequence similarity and genetic distance values for material under these two names are indicative of the existence of a single species (0.006 ± 0.002; 99.73 ± 0.10%); hence, the distinction of these two entities cannot be supported by the use of this marker alone. However, G. destructans and G. dunense are distinguished following the outcome of a multigene analysis (Tchotet Tchoumi et al. 2018).

G. steyaertanum B.J. Smith & Sivasith. forms a well-supported group (89%, 0.99; Fig. 5) which is composed of 39 entries from specimens growing on eudicots in Indonesia and Australia (Table 1). G. steyaertanum was proposed as the correct name for the erroneously-labelled “G. lucidum” specimens reported to occur in this particular region (Smith and Sivasithamparam 2003). However, none of the sequences hereby examined was originally deposited as “G. lucidum”; instead, entries were labelled either as G. steyaertanum (34) or as G. aff. steyaertanum (3). This species forms a sister clade (98%, 1.00; Fig. 5) with G. mizoramense Zothanzama, Blanchette, Held, C.W. Barnes represented by only three identical sequences, all deriving from India (Suppl. material 1: Table S2 and Suppl. material 5: Fig. S2c). The two taxa appear closely related as evidenced by their genetic distance and sequence similarity values (0.016 ± 0.005 and 98.46 ± 0.31%, respectively), while the respective sequences differ at six conserved positions (Suppl. material 3: Fig. S1); furthermore, three conserved nucleotides in ITS2 are common to both and separate them from other related taxa of Cluster A.3. The very limited representation of G. mizoramense does not allow any definite conclusions regarding its taxonomic status.

G. martinicense Welti & Courtec. is sister to G. multipileum (72%, 0.98; Fig. 5), the two taxa being closely related (98.76 ± 0.36% and 0.012 ± 0.003) with no barcoding gap existing between them (Fig. 8). The former (93%, 1.00; Fig. 5) consists of 49 entries originating from specimens collected in USA, Mexico, Cuba, Martinique, Colombia, Brazil and Argentina which were deposited under various names, i.e. G. martinicense (18 entries, including type material), “G. parvulum” (24), “G. perzonatum” (2), “G. tuberculosum” (1), “G. oerstedii” (1) and “G. lucidum” (1) (Table 1, Suppl. material 5: Fig. S2c). Although at least three (not adequately supported) groups are evident in the expanded tree (which corresponds to different geographic origins, i.e. Colombia, Brazil/Argentina and Cuba/Martinique; Suppl. material 5: Fig. S2c), ‘intergroup’ values for ITS sequence divergence do not justify the existence of more than one phylogenetic species. Therefore and until more information becomes available, we prefer to maintain them all under G. martinicense. The name “G. parvulum” corresponds to a taxon forming part of Cluster A.2, as previously discussed.

G. multipileum Hou is phylogenetically supported (74%, 0.97; Fig. 5) and hereby represented by 243 entries; 130 entries are grouped in 29 ASVs (Suppl. material 1: Tables S2, S4). The respective collections originate from south and east Asia on a wide range of eudicot (mostly) and monocot or gymnosperm hosts, including also commercial strains (Suppl. material 5: Fig. S2c). The majority of sequences were labelled either as “G. lucidum” (105) or as Ganoderma sp. (112) and only 22 were identified as G. multipileum.

A second group (A.3.2; 100%, 1.00; Fig. 5) within Cluster A.3 includes sequences stemming from south and ast Asia. G. tropicum (Jungh.) Bres. is a species described from Indonesia (Java) and is widely distributed across subtropical and tropical Asia (Steyaert 1972; Moncalvo and Ryvarden 1997; Luangharn et al. 2019a). According to Corner (1983), this is a complex of pantropical occurrence, comprising many taxonomic varieties characterised by strongly echinulate basidiospores. However, pertinent sequenced material derives only from Asia. Hence, 15 entries, corresponding to specimens from China, Laos, Taiwan and Thailand, were originally identified as G. tropicum (type material included; Table 1). Moreover, 12 entries, labelled as “G. fornicatum” from specimens collected in India and Taiwan, presented genetic distances and sequence similarity indicative of a high affinity to G. tropicum sequences (i.e. 0.010 ± 0.004; 98.93 ± 0.63%, respectively). The type material of G. fornicatum (Fr.) Pat. originates from Brazil, but cannot be located and is most probably lost (Ryvarden 1991; Moncalvo and Ryvarden 1997). Furthermore, no recent information exists on the presence of this species in the Neotropics (Wang et al. 2014) and no sequence is available from specimens deriving from this particular region. In contrast, sequences/material under this name originates from Asia only (Imazeki 1939; Zhao and Zhang 2000; Wang and Wu 2008). Last, according to Mycobank, the current name for G. fornicatum is G. orbiforme (Fr.) Ryvarden. However, the latter is positioned in Clade D of the present analysis and is composed of entries originating solely from Brazil (Fig. 5). In view of the above, we maintain those particular entries identified as “G. fornicatum” under G. tropicum.

A strongly-supported terminal subclade in Group A.3.2 (100%, 1.00; Fig. 5) consists of 15 sequences (five ASVs) initially labelled “G. tropicum”, originating from Indian specimens obtained from various angiosperms, for example, Ficus benghalensis, Terminalia bellirica, Delonix regia and Cassia fistula (Arulpandi and Kalaichelvan, unpublished results). It is considered as a new phylogenetic species, hereby referred to as “Ganoderma sp. A6” (corresponding to the UNITE DOI SH1723103.08FU) since it is clearly separated from G. tropicum by presenting interspecific genetic distance and sequence similarity values of 0.026 ± 0.005 and 97.24 ± 0.68%, respectively.

Furthermore, three singletons from Malaysian material (in this particular case, geographic origin is inferred from the title of the study which appears on the respective GenBank records), initially identified as “G. fornicatum”, form another distinct well-supported group (100%, 1.00; Fig. 5). This is also considered as a new phylogenetic species provisionally named “Ganoderma sp. A7” (corresponding to the UNITE DOI SH1723183.08FU) and is well distinguished from G. tropicum which is the closest taxon amongst those examined (sequence similarity: 94.64 ± 1.56%; genetic distance: 0.046 ± 0.015).

Cluster D.1 is placed at the base of Clade D and it is strongly supported (98%, 1.00; Fig. 6). It consists of five sequences corresponding to eight entries; two entries are under the name G. mbrekobenum E.C. Otto, Blanchette, Held, C.W. Barnes & Obodai from Ghana (including the type material representing a laccate stipitate taxon; Otto et al. 2016), while the rest originated from Asia (India and Sri Lanka; Fig. 6) on a broad range of plant hosts and were deposited as Ganoderma sp. Similarly, good support for Cluster D.1 was obtained when additional (23) singletons/sequences were included in the analysis (99%, 1.00; Suppl. material 7: Fig. S2e, Suppl. material 1: Table S2); this larger sample-set presented intraspecies sequence similarity of 98.65 ± 0.70%. Although two or more subgroups are formed within this species (Fig. 6 and Suppl. material 7: Fig. S2e), ITS similarity and genetic distance values do not adequately support their distinct status; therefore and in the absence of additional data, we prefer to maintain a single species (i.e. G. mbrekobenum) in Cluster D.1.

Cluster D.2 comprises two distantly-related species, i.e. G. sinense Zhao, Hsu & Zhang (Zhao et al. 1979b) and G. nasalanense Hapuar., Pheng. & K.D. Hyde (Hapuarachchi et al. 2019). The monophyletic G. sinense (100%, 1.00; Fig. 6) includes 26 sequences representing 66 entries deriving from Chinese material plus four collections from Thailand and a single collection from Taiwan (Suppl. material 1: Table S2). Most of them were labelled as G. sinense (45), while six were deposited under the names “G. japonicum”, “G. formosanum” and “G. atrum”. This finding is in accordance with reports advocating synonymy of G. sinense and G. japonicum (Fr.) Sawada (Liao et al. 2015; Wang and Yao 2005), whose authors examined specimens from China only. Therefore, in the absence of material from the type locality (Japan), a definite conclusion cannot be drawn regarding the status of G. japonicum. In addition, the sole sequence available under the name G. atrum Zhao, Hsu & Zhang (JQ886403; China, Hainan Island) grouped together with G. sinense in the present phylogeny. Similarly, a sequence (the joinder of X78752 and X78773) from a specimen originally identified as G. formosanum T.T. Chang & T. Chen was also positioned within G. sinense. On the other hand, G. nasalanense is strongly supported (98%, 1.00; Fig. 6) and includes two identical sequences (one from the type material) from Laos plus 13 entries from specimens collected in India, Malaysia and Vietnam, initially labelled as “G. australe”, Ganoderma sp. and “uncultured soil fungus” (Table 1, Suppl. material 7: Fig. S2e). Sequence similarity and genetic distance values support the distinct phylogenetic status of G. nasalanense with respect to G. sinense as evidenced by the clear barcoding gap they exhibit (94.17 ± 0.55% and 0.051 ± 0.002, respectively; Fig. 8).

Cluster D.3 is well supported (75%, 1.00; Fig. 6) and includes four species. One corresponds to G. cupreum (Cooke) Bres. described on the basis of material collected in Africa and is represented by eight entries; three of them originate from Cameroon, one (deposited as “G. cf. cupreum”) from South Africa, one as “G. chalceum” (originating from Tanzania on the basis of the respective submission’s title), one environmental sample from Gabon and one of unknown origin, while a “G. australe” entry from Malaysia was placed at the base of the subclade. We maintain the name G. cupreum (since it has priority over G. chalceum) for the respective phylogenetic species, which forms a well-supported clade (99%, 1.00; Fig. 6) presenting high intraspecific values for sequence similarity (99.61 ± 0.21%) and low for genetic distance (0.004 ± 0.002).

Another terminal subclade is composed of nine entries (98%, 1.00; Fig. 6); five of them were originally deposited as G. ecuadoriense W.A. Salazar, C.W. Barnes & Ordóñez (three from Ecuador, including the type specimen and two from Brazil), one as G. subfornicatum Murrill, two as Ganoderma sp. (Brazil and Peru) and one labelled as “uncultured fungus” (Suppl. material 1: Table S2). The latter is the only one not originating from the Neotropics and derives from soil sampled in India (KJ411557). Material labelled as G. subfornicatum was identified as G. orbiforme (Fr.) Ryvarden (Ryvarden and de Meijer 2002) and, hence, the former was not included in a subsequent study of Ryvarden (2004) on neotropical polypores. On the other hand, Torres-Torres et al. (2012) treated G. subfornicatum and G. orbiforme as distinct taxa despite their being very similar in morphology. In addition and in the context of the present study, it was evidenced that the sole sequence available under the name G. subfornicatum (JX082352, French Guiana) was placed within the same terminal clade as sequence KU128524 representing the holotype of G. ecuadoriense. The latter is a recently-described species from Ecuador (Salazar et al. 2016); however, although those authors conducted a morphological and molecular study on their material, they did not include specimens of G. subfornicatum, which commonly occurs in the same larger area (type material from Belize) (Welti and Courtecuisse 2010; Torres-Torres et al. 2012). On the basis of the previously-presented information, G. subfornicatum is maintained to describe this terminal subclade (although only one sequence is available under this name in GenBank) and “G. ecuadoriense” is abandoned as nom. illeg.

Five sequences, representing G. orbiforme, form a well-supported terminal group (89%, 0.96; Fig. 6) which, however, presents high affinity to the sister clade of G. subfornicatum (100%, 0.96; Fig. 6), as is evidenced by the absence of a barcoding gap in the pairwise comparison of the two taxa (genetic distance and sequence similarity: 0.011 ± 0.002 and 98.69 ± 0.23%, respectively). G. orbiforme was originally described from Guinea, but was also reported from South America (Ryvarden 2000; Baltazar and Gibertoni 2009; Gomes-Silva et al. 2011; Torres-Torres et al. 2012). Although none of these studies included molecular data, in some of them (e.g. Ryvarden 2000; Torres-Torres et al. 2012), the type specimen of G. orbiforme was examined alongside Ganoderma basidiomes from the Neotropics to assess the identity of the latter. Moreover, four sequences from Brazilian specimens identified as G. orbiforme (de Lima Júnior et al. 2014) were grouped within this particular subclade. Therefore, we believe that G. orbiforme is the correct name for this phylogenetic species which occurs in the Neotropics. The use of this name for material originating from Asia and Australia (Wang et al. 2014; Hapuarachchi et al. 2019) is not supported by molecular evidence and erroneous conclusions could be attributed to the high morphological variability of specimens belonging to closely-related taxa of Cluster D.3, for example, G. mastoporum and G. orbiforme (Fig. 6).

Finally, a well-represented and supported terminal subclade of Cluster D.3 (80%, 1.00; Fig. 6) includes 122 entries (95 of them are singletons; Suppl. material 1: Table S2) deriving from specimens collected in southeast Asia and Australia mainly on eudicot angiosperms. It consists of sequences deposited under the names G. mastoporum (12), “G. orbiforme” (19), “G. cupreum” (10), “G. fornicatum” (3), “G. multicornum” (1) and Ganoderma sp. (11), as well as under “G. australe” (60, resulting from a single study conducted on material originating from Borneo). On the basis of what was previously mentioned about the correct phylogenetic position of G. orbiforme and although sequences labelled as “G. orbiforme” deriving from China were included under this name in previous studies (Wang et al. 2014; Hapuarachchi et al. 2018b; Xing et al. 2018), we believe that relevant material originating from southeast Asia and Oceania corresponds to G. mastoporum (Lév.) Pat. (initially described on the basis of material collected in Singapore) and that the 12 G. mastoporum entries in this particular subclade of Cluster D.3 were correctly identified. Therefore, G. mastoporum cannot be considered as a synomym of G. orbiforme (as stated in MycoBank and Index Fungorum) since the latter name corresponds to a related – yet distinct – phylogenetic species (interspecific sequence similarity and genetic distance values: 97.34 ± 0.49% and 0.025 ± 0.005, respectively). Similarly, G. mastoporum and G. cupreum could be separated on the basis of the outcome of the present study because they form distinct well-supported terminal groups (Fig. 6, Suppl. material 7: Fig. S2e) presenting interspecific sequence similarity of 97.20 ± 0.45%. One entry, labelled as G. multicornum (MT772000, Suppl. material 1: Table S2; Suppl. material 7: Fig. S2e), shows relatively high affinity (genetic distance: 0.012 ± 0.008, sequence similarity: 98.65 ± 0.70%) with the rest of the entries within the G. mastoporum clade (Fig. 6, Suppl. material 7: Fig. S2e).

Cluster D.4 is strongly supported (93%, 1.00; Fig. 6) and is further divided into two sister groups, i.e. Group D.4.1 and D.4.2; the former includes species collected on angiosperms and gymnosperms, while the latter, the so-called “palm group”, comprises sequences from material mostly associated with monocotyledons.

Group D.4.1 (87%, 1.00; Fig. 6) includes the recently-introduced sessile and laccate G. angustisporum J.H. Xing, B.K. Cui & Y.C. Dai reported from China on Casuarina equisetifolia (Xing et al. 2018). Moreover, sequences from specimens of different origin (India, Malaysia and Australia, labelled as “G. australe” and Ganoderma sp.; Table 1) growing on other eudicots also nested together with G. angustisporum and present high intraspecific similarity values (99.07 ± 0.43%), thus largely expanding the known distribution of this species and the number of host plant taxa it is associated with (Table 1, Suppl. material 7: Fig. S2e). A closely-related entity to G. angustisporum is found as a terminal subgroup (100%, 1.00; Fig. 6) and it includes two singletons initially labelled as “G. applanatum” from material collected in Africa (Gabon) on eudicots. It presents genetic distances and sequence similarity values of 96.87 ± 1.21% and 0.027 ± 0.004, respectively, vs. G. angustisporum. It is therefore considered as a new phylogenetic species and is provisionally labelled as “Ganoderma sp. D1” (corresponding to the UNITE DOIs SH1740449.08FU and SH1740450.08FU).

Group D.4.2 (91%, 1.00; Fig. 6) is composed of five species comprising material originating mainly from monocot angiosperms. Three of them are distinguished by a clear barcoding gap (Fig. 8) and form a well-supported terminal subclade (79%, 0.99, Fig. 6). G. zonatum Murill (100%, 1.00; Fig. 6 and Suppl. material 7: Fig. S2e) consists of 84 correctly-labelled entries with high similarity values (99.86 ± 0.08%), originating from specimens growing on palms in southeast USA. G. ryvardenii Tonjock & Mih (100%, 1.00; Fig. 6) is represented by 22 entries mainly deriving from Cameroon (including the sequence from the type material, Table 1). Moreover, four entries, labelled as Ganoderma sp. from Colombia, form a distinct strongly-supported clade (100%, 1.00; Suppl. material 7: Fig. S2e) representing, thus, a new phylogenetic species provisionaly named Ganoderma sp. D2 (corresponding to the UNITE DOI SH1723113.08FU).

The other two species comprise material from Asia only; the ‘core’ part corresponds to G. boninense Pat. and it is represented by 69 entries deposited as G. boninense (32), “G. miniatocinctum” (3), “G. orbiforme” (2), “G. zonatum” (3) and Ganoderma sp. (29) (Table 1; Fig. 6). The rest of the entries form a terminal subgroup with strong support (98%, 1.00; Fig. 6) and includes material originating from Indonesia. This material was originally identified to genus level only (12 entries, labelled as Ganoderma sp.; Table 1). Sequence similarity and genetic distance values vs. G. boninense (97.30 ± 0.67% and 0.029 ± 0.006, respectively) are indicative of the presence of a new phylogenetic species, provisionally named Ganoderma sp. D3 (corresponding to the UNITE DOIs SH1723050.08FU and SH1723098.08FU).

On the basis of the results presented above, it is apparent that G. zonatum is not related to G. sessile (the latter forms part of Cluster A.2), as previously reported by Gottlieb et al. (2000); the former name was misapplied to a specimen originating from Argentina. Neither G. zonatum nor G. boninense forms part of the G. lucidum complex (Zhou et al. 2015) since they are distinctly positioned into Clade D (Fig. 4). In addition, three sequences available under the name G. miniatocinctum Steyaert are grouped together with G. boninense material.

Cluster E.1 corresponds to a single strongly-supported phylospecies (99%, 1.00; Fig. 7) distinctly placed at the base of Clade E. It includes material from tropical/subtropical Asia which is mainly associated with the name “G. australe” (29 entries) (Table 1); however, G. williamsianum Murrill. (7) is the correct name to assign to this terminal group since relevant descriptions and reported occurrence (Steyaert 1972; Corner 1983; Wang and Wu 2010; Xing et al. 2018) are in agreement with the phylogenetic position presented here. In addition, G. williamsianum presents high intraspecific similarity values (99.47 ± 0.22%) despite the large number of entries examined and their rather wide geographic origin. Cluster E.1 corresponds to ‘clade 8’ of the phylogeny presented by Moncalvo and Buchanan (2008).

Cluster E.2 (84%, 0.99; Fig. 7) includes 43 sequences representing 176 entries (or 177 sequences deriving from 310 entries in the expanded dataset; Suppl. material 8: Fig. S2f). It consists of seven phylogenetic species, two of which have a neotropical distribution (including USA) while the rest occur in Asia, Africa and Oceania (Table 1).

A well-supported terminal subclade (87%, 0.99; Fig. 7) is formed by 23 entries with high intraspecific sequence similarity (99.64 ± 0.23%, min. 99.11%) deriving from material originating from the Neotropics on eudicots under various names, i.e. “G. tornatum” (6), “G. lobatum” (3), “G. parvulum” (1), “G. gibbosum” (1), “G. applanatum complex” (8) and Ganoderma sp. (4) (Table 1 and Suppl. material 1: Table S2). Since some of these names are in use for other sequences examined in this study (positioned in other Clades/Clusters, for example, as is the case for G. parvulum and G. lobatum), while for others there is not adequate evidence to support their correct use in this particular case (e.g. G. tornatum), we prefer to label this distinct phylogenetic entity as Ganoderma sp. E1 (corresponding to the UNITE DOI SH1723047.08FU). This species corresponds to ‘clade 7’ in the study of Moncalvo and Buchanan (2008). As is the case elsewhere in the present study, properly/accurately identified type material (representing one or more of the taxa whose names appear in this subclade) needs to be sequenced in order to arrive at robust conclusions regarding the real identity of this particular phylospecies.

A strongly-supported (96%, 1.00; Fig. 7) sister group to the previous phylospecies consists of 37 sequences deriving from specimens collected in South and Central America, as well as in the USA (Florida) on a wide range of host-plants, for example, Cenostigma pluviosum var. peltophoroides, Inga vera, Jacaranda mimosifolia, Leucaena leucocephala and Elaeis guineensis. These sequences, initially deposited under various names (i.e. “G. gibbosum” (12), “G. tornatum” (8), “G. lobatum” (7), “G. australe” (2) and “G. applanatum complex” (2)), are hereby placed under the name Ganoderma sp. E2 (corresponding to the UNITE DOI SH1723047.08FU). We believe that accurate association of this species to any established Ganoderma taxon name is not possible until additional information – through the study of relevant type material – becomes available. Although Ganoderma sp. E1 and Ganoderma sp. E2 demonstrate relatively-high sequence similarity and rather low genetic distances in pairwise comparisons (97.73 ± 0.54% and 0.017 ± 0.005, respectively), there is no overlap between the respective intra- and interspecies values; their distinct species status is therefore proposed.

Cluster E.2 also includes a large terminal subclade (65%, 1.00; Suppl. material 8: Fig. S2f) comprising material from Asia, Africa and Oceania. Clade E.2 corresponds to the G. gibbosum complex and consists of (at least) five species, i.e. G. gibbosum (Blume & T. Nees) Pat., the recently introduced G. eickeri Tchotet, M.P.A. Coetzee, Rajchenb. & Jol. Roux, G. ellipsoideum Hapuar., T.C. Wen & K.D. Hyde, as well as two new phylospecies (Ganoderma sp. E3 and Ganoderma sp. E4). G. gibbosum is composed of 107 entries (40 of which are singletons) labelled mainly under the names G. gibbosum (59), “G. applanatum” (27) and “G. australe” (8) (Suppl. material 1: Table S2 and Suppl. material 8: Fig. S2f). It shows low intraspecific sequence variability (sequence similarity: 99.58 ± 0.17%; genetic distance: 0.004 ± 0.002). A distinct entity, here called G. aff. gibbosum, is exclusively composed of sequences of Indian origin (51 entries including 10 singletons; 69%, 1.00; Fig. 7) which are mainly deposited as Ganoderma sp. However, the relatively-high affinity exhibited by sequences of G. aff. gibbosum vs. other G. gibbosum sequences from Asia (sequence similarity and genetic distance values: 99.04 ± 0.27% and 0.009 ± 0.002, respectively) prevents us from considering it as a distinct phylospecies, at least until further evidence becomes available.

The other terminal clade (92%, 1.00; Fig. 7) represents the recently-introduced G. eickeri consisting of four entries, including the type material, which originate from South Africa on angiosperms (Tchotet Tchoumi et al. 2019). However, on the basis of the ITS data evaluated in this study, G. eickeri appears closely related to G. gibbosum (interspecific sequence similarity and genetic distance values: 98.90 ± 0.25% and 0.011 ± 0.003, respectively).

G. ellipsoideum is a recently-described species from Hainan Island, China (Hapuarachchi et al. 2018b). It is here represented by 81 entries corresponding to 10 ASVs and 44 singletons (0.99; Fig. 7). Sequences within the G. ellipsoideum subclade (including the one from the type material) were mainly labelled as “G. gibbosum” (22), “G. australe” (10), “G. australe cplx” (10), “G. applanatum” (5) and Ganoderma sp. (10); they originate from material of broad geographic distribution (south and east Asia, Oceania) on eudicots. In addition, this study revealed that two environmental samples deriving from the USA (UDB0769802 and UDB0763546) formed part of this terminal subclade. As in the case of G. eickeri, G. ellipsoideum is closely related to G. gibbosum (sequence similarity and genetic distance values: 98.84 ± 0.19% and 0.012 ± 0.002, respectively) and no barcoding gaps were detected with respect to the closest related species (i.e. G. gibbosum and G. eickeri; Fig. 8); therefore, its distinct phylogenetic status is questionable on the basis of ITS data. G. gibbosum and G. ellipsoideum correspond to ‘clade 5’ of the phylogeny presented by Moncalvo and Buchanan (2008).

The other two phylogenetic species appearing on the terminal subclade of Cluster E.2 (96%, 1.00; Fig. 7) are hereby designated as Ganoderma sp. E3 (corresponding to the UNITE DOIs SH1723116.08FU and SH1723270.08FU) and Ganoderma sp. E4 (corresponding to the UNITE DOI SH1677211.08FU); both demonstrate high intraspecific sequence similarity values (> 99.31%). The former consists of six sequences deposited as “G. australe” deriving from Indonesian and Australian material. The latter species (100%, 1.00; Fig. 7) corresponds to ‘clade 6’ in the study of Moncalvo and Buchanan (2008) and includes 13 entries from specimens originating from Malaysia and Indonesia on eudicots, which were also labelled “G. australe” (Table 1). Both species are well separated from each other (Ganoderma sp. E3 vs. Ganoderma sp. E4: 94.67 ± 1.02% and 0.030 ± 0.006 for sequence similarity and genetic distance values, respectively), as well as from the rest of the species within Cluster E.2, for example, Ganoderma sp. E3 vs. G. ellipsoideum: 96.84 ± 1.01% and 0.025 ± 0.006, respectively.

Cluster E.3 consists of material corresponding to the laccate taxa G. pfeifferi Bres. (17 sequences from Europe only; Table 1), G. cupreolaccatum Kalchbr. ex Z. Igmándy (invalid name, type locality: Hungary), G. mutabile Cao & Yuan (two sequences from China, including one from the type specimen) and the recently-introduced G. knysnamense Tchotet, M.P.A. Coetzee, Rajchenb. & Jol. Roux (four entries from South Africa, including the type; Tchotet Tchoumi et al. 2019), which were adequately supported (71%, 0.97; Fig. 7). Amongst them, G. knysnamense is placed at the base of this cluster (100%, 1.00; Fig. 7) and is well separated from the closest species (i.e. G. mutabile: 0.036 ± 0.001 and 96.43 ± 0.13%). Furthermore, G. pfeifferi and G. mutabile are placed on well-supported terminal subclades (90%, 1.00; Fig. 7) and they both present high intraspecific similarity (> 99.84%); however, the respective interspecific values in pairwise comparisons are indicative of their affinity (98.14 ± 0.19%, 0.019 ± 0.005; Fig. 8). G. mutabile was introduced as a distinct species after examining only one collection which, according to its authors (Cao and Yuan 2013), resembles G. pfeifferi with respect to the dark brown context and the similar spore size; however, it differs by its laccate crust and non-stratified tubes. The other noteworthy sequence within Cluster E.3 derives from a specimen identified as G. cupreolaccatum, which was previously considered as a facultative (heterotypic) synonym of G. pfeifferi (Mycobank). However, both similarity and genetic distance values of this particular sequence vs. G. pfeifferi are indicative of a distinct species (i.e. 94.97 ± 0.17% and 0.030 ± 0.007, respectively). Still, the variability of this unique sequence is located exclusively at the beginning of ITS1, a region generally conserved between closely-related taxa of the genus (Nilsson et al. 2017), while the rest is identical to those of G. pfeifferi. Therefore, questions are raised about the quality of this particular sequence and no definite conclusions could be drawn concerning the exact status of G. cupreolaccatum.

The close phylogenetic position of G. pfeifferi and G. adspersum (Schulzer) Donk (Cluster E.5) is congruent with their similar pileus dark-brown context and the complex structure of the crust in contrast to the normal palisade appearance in laccate species of Clade A (Hansen 1958; our observations); still, G. pfeifferi clearly differs by the laccate pileus and the width and quotient of spores (Ryvarden and Gilbertson 1993; Niemelä and Miettinen 2008; Ryvarden and Melo 2017). In addition, these characteristics, in conjunction with the cracked/wrinkled resinous layer on the pileus, distinguish G. pfeifferi from specimens of G. lucidum and G. resinaceum (Ryvarden and Gilbertson 1993). Furthermore, the outcome of the present study evidences the grouping of G. pfeifferi within Clade E and contrasts previous reports by Hseu et al. (1996) and Cao and Yuan (2013), which linked this taxon with the G. lucidum complex and G. resinaceum, respectively. This discrepancy is apparently due to an initial misidentification of strain CBS 747.84 (sequences X78738/ X78759) which was labelled as “G. pfeifferi” instead of G. resinaceum (Moncalvo et al. 1995a).

Cluster E.4 (0.99; Fig. 7, and 82%, 1.00; Suppl. material 8: Fig. S2f) includes five species represented by sequences from material with a world-wide distribution (apart from Europe) occurring on angiosperms. A well-supported subgroup within this cluster is formed by the recently-introduced G. chocoense J.A. Flores, C.W. Barnes & Ordoñez, including only the sequence from the type material (Ecuador; Flores et al. 2018) and G. podocarpense J.A. Flores, C.W. Barnes & Ordoñez (Flores et al. 2017), consisting of two sequences from Ecuador and Panama. Despite their overlapping distribution range, the values of sequence similarity (97.57 ± 0.13%) and genetic distance (0.025 ± 0.003) are indicative of their distinct status at species level.

A sister group to the former (1.00; Fig. 7 and 95%, 1.00; Suppl. material 8: Fig. S2f) is composed of eight sequences originating from Argentinian material under the names “G. lobatum”, “G. tornatum” and Ganoderma sp. (Table 1). On the basis of the information available, it is not possible to assign a particular taxonomic name to these sequences. Therefore, we prefer to designate this phylogenetic species as Ganoderma sp. E5 (corresponding to the UNITE DOI SH1678465.08FU) (100%, 1.00; Fig. 7). The latter entity is well separated from G. podocarpense, i.e. sequence similarity: 95.58 ± 0.59% and genetic distance: 0.044 ± 0.006.

A sister terminal subclade to the previous group (G. chocoense, G. podocarpense and Ganoderma sp. E5) consists of sequences originating from material collected in south and southeast Asia, while a single sequence indicates its presence also in Papua New Guinea (95%, 1.00; Fig. 7). They are labelled mainly as “G. australe” (19), “G. applanatum” (2) and Ganoderma sp. (8) (Table 1). Again, none of the initially-assigned names could be maintained; therefore, we refer to this distinct phylogenetic species as Ganoderma sp. E6 (corresponding to the UNITE DOI SH1723070.08FU) presenting low interspecific sequence similarity and high genetic distance vs. Ganoderma sp. E5 (i.e. 95.59 ± 0.70% and 0.040 ± 0.006, respectively). Ganoderma sp. E5 and Ganoderma sp. E6 appear as sister subclades in ‘clade 4’ of the phylogeny presented by Moncalvo and Buchanan (2008). Moreover, a single sequence (AF255183; Suppl. material 1: Table S2 and Suppl. material 8: Fig. S2f), originating from New Zealand, seems to correspond to a distinct entity, which however does not fulfil the criteria set in this study to merit recognition as a Ganoderma phylospecies (genetic distance and sequence similarity values vs. Ganoderma sp. E6: 0.017 ± 0.002 and 98.27 ± 0.18, respectively).

Finally, a distinct group included 80 entries deriving from specimens mainly from southeast Asia, South America and Oceania, as well as from central/north America and South Africa. That said, one sequence originated in the vicinity of the Kew Botanical Gardens, UK, such that we suspect it to have been imported with plant material. These sequences were mostly deposited as “G. australe” (27), “G. australe complex” (14) and Ganoderma sp. (13). Despite the widespread use of the “australe” epithet to describe several entities (14) placed in other terminal clades of the present study, we believe that this particular terminal group coincides with G. australe (Fr.) Pat. after evaluating all available ITS sequence data, the geographic distribution of specimens analysed and pertinent publications (Ryvarden and Gilbertson 1993; Moncalvo and Ryvarden 1997; Moncalvo and Buchanan 2008). As regards sequence similarity, intraspecific values are notably high (99.56 ± 0.25%) considering the diverse geographic origin of the respective material, whereas interspecific values to the closest taxon is low (i.e. Ganoderma sp. E6; 95.04 ± 0.66%). Although ambiguities exist about the exact distribution range of G. australe (Yeh and Chen 1990; Buchanan and Wilkie 1995; Moncalvo and Ryvarden 1997; Smith and Sivasithamparam 2000) and the type specimen originating from the Pacific area is lost, this appears to be the most common Ganoderma species throughout the tropics and subtropics and corresponds to ‘clade 3’ in the study of Moncalvo and Buchanan (2008).

Cluster E.5 (97%, 1.00; Fig. 7) comprises sequences from specimens originating from the Northern Hemisphere and it is further divided into two well-supported subclades. One corresponds to a new phylogenetic species hereby designated as Ganoderma sp. E7 (corresponding to the UNITE DOI SH1723077.08FU), which is composed of 17 entries (100%, 1.00; Fig. 7) deposited under the name “Ganoderma sp.”, “G. lobatum” and “G. applanatum cplx” deriving from US material only (intraspecific sequence similarity: 98.83 ± 0.90%; interspecific sequence similarity vs. G. adspersum: 95.29 ± 0.51%). “Ganoderma sp. E7” forms part of ‘clade 2’ in the phylogeny presented by Moncalvo and Buchanan (2008).

A sister subclade to the aforementioned clade corresponds to G. adspersum sensu lato and is represented by 155 entries deriving from Europe, south and west Asia and North Africa (1.00; Suppl. material 8: Fig. S2f). G. adspersum (Schulzer) Donk is a common species in the Palearctic realm; in Europe, it is usually reported on a wide range of angiosperms (Ryvarden and Gilbertson 1993), but it also appears on gymnosperms (i.e. Abies cephalonica; Zervakis et al. 1998; this study). Most of the sequences (117) examined in this study were properly identified as G. adspersum. However, several misidentifications were noted, mainly for Asian material labelled as “G. applanatum”, “G. australe” and “G. australe complex” (Suppl. material 1: Table S2). In fact, G. adspersum was considered as a synonym of G. australe (Ryvarden 1976; Ryvarden and Gilbertson 1993; Ryvarden and Melo 2017), but ITS sequence data clearly separated it from G. australe. The outcome of this study reveals that G. adspersum includes specimens with a distribution ranging from Europe to India (89%, 1.00; Fig. 7). Moreover, nine entries from samples, originally identified as G. adspersum, “G. australe complex”, “G. applanatum” and “uncultured fungus” (China, Japan and South Korea; Suppl. material 1: Table S2), form a distinct terminal subgroup (72%; Fig. 7), which is hereby named G. aff. adspersum on the basis of its close affinity to G. adspersum (i.e. sequence similarity and genetic distance values: 98.47 ± 0.33% and 0.013 ± 0.004, respectively).

In conclusion, the plasticity of morphological characters and substantial overlap of alleged diagnostic features is prevalent in the ‘dull’ taxa of this group; consequently, their taxonomic significance is dubious. Moreover, the massive, heavily-agglutinated matrix of the pileal crust in such non-laccate species often obstructs observation of discriminating features in pileal elements. The situation is further aggravated by the loss of type material of widely and commonly used species names (e.g. G. australe and G. gibbosum), the absence of (correct) neo-typification (as there are ambiguous synonymies regarding several taxa) and unclear species descriptions (e.g. G. tornatum and G. lobatum). All of the above could explain the obstacles which prevented the establishment of a stable classification system for species of Clade E. Still, the highly variable morphology of basidiomes and the ITS divergence are indicative of underestimated diversity and the presence of cryptic species is quite certain as is also indicated by the outcome of the present study.