Marine sediments examined in this study were collected during several surveys conducted along the coast of Catalonia (Spain) between 2021 and 2023. Sampling sites included the area in front of two beaches on the Tarragona coast (Miracle and Arrabassada), located near the discharge area of the Francolí River, and the discharge areas of the rivers Ebro (Tarragona Province), Ter (Girona Province), and Llobregat (Barcelona Province) (Fig. 1). The latter three locations are included in the natural parks of the Ebro Delta, the Medas Islands, and the Llobregat Delta, respectively.

At each site, samples were collected in sets of two independent surveys, and sediments were recovered from around 6, 13, 20, and 27 m depth. The Llobregat River site was an exception, with only one survey conducted, during which samples were collected from 20 and 30 m depths. At each depth, four sub-samples were taken from 15 cm below the seabed surface, using 50 mL plastic tubes. Throughout the collection process, tubes were stored in a refrigerated container until their arrival at the laboratory.

At each area, sediment samples from different depths were treated separately. However, the sub-samples from each depth were mixed in a container, vigorously shaken, and poured onto a plastic tray with several layers of sterile filter paper to remove excess water. Subsequently, each mixed sample was assessed through both direct plating and an SMF pre-treatment, following the procedure explained in Guerra-Mateo et al. (2024).

The culture media used for isolation were dichloran rose bengal chloramphenicol agar (DRBC; 5 g peptone, 10 g glucose, 1 g KH₂PO₄, 0.5 g MgSO₄, 25 mg rose bengal, 200 mg chloramphenicol, 2 mg dichloran, 15 g agar, 1 L distilled water); potato dextrose agar (PDA; Condalab, Madrid, Spain) supplemented with 2 g/L of cycloheximide (PDA+C); and two additional media made with seawater, 3% malt extract agar (SWMEA3%; 30 g malt extract, 5 g mycological peptone, 15 g agar, 1 L seawater) and oatmeal agar (SWOA; 30 g oatmeal, 15 g agar, 1 L seawater). To inhibit bacterial growth, 5 mL of chloramphenicol (15 g/L ethanol) was added to SWMEA3%, SWOA, and PDA+C media.

For direct plating, 1 g of sediment from each sample was distributed across two Petri dishes per medium and mixed with either melted PDA+C, SWMEA3%, or SWOA at 45 °C. In the case of DRBC, only 0.5 g of sediment was distributed across two Petri dishes to address the fast-growing fungi. A similar methodology was applied to the sediment pre-treated with SMF, with the exception that 1 mL of floccule from each sample was distributed across two Petri dishes for PDA+C, SWMEA3%, and SWOA, and 0.5 mL of floccule was mixed with DRBC. Each sample was cultured in duplicate for both approaches; one set of plates with different media was incubated at 22–24 °C, and the other at 15 °C. All cultures were incubated in darkness and examined weekly under a dissecting microscope for up to 5–8 weeks.

Pure cultures were obtained on PDA from fragments of the colonies or conidia in the primary plates using a sterile dissection needle. These cultures were used to provide a preliminary morphological identification. Strains that did not exhibit sporulation were further subcultured to potato carrot agar (PCA; 20 g potato, 20 g carrot, 15 g agar, 1 L distilled water) or oatmeal agar (OA; 30 g oatmeal, 15 g agar, 1 L distilled water) to promote sporulation.

The strains obtained were morphologically examined for a presumptive generic identification. A maximum of three identical strains per sample were selected for sequencing based on the conserved morphological traits frequently displayed between phylogenetically close ascomycetous species (Gannibal 2022). The strains included in this study represented a selection of the onygenalean fungi recovered from the marine sediment samples (Table 1). Identification of the strains at the species level was based on a polyphasic approach under the consolidated species concept, combining multi-locus phylogenetic analyses, morphology, and ecological data (Quaedvlieg et al. 2014). Metadata about the identified strains has been deposited in the Global Biodiversity Information Facility (GBIF, https://doi.org/10.15468/emtvnj, accessed in July 2025).

Genomic DNA was extracted using a chloroform/isopropanol protocol (Tripathy and Maharana 2017) and quantified with a Nanodrop 2000 (Thermo Scientific, Madrid, Spain). We amplified the ITS region and the LSU D1–D3 domain of the nrDNA using the primer pair ITS5/LR5 (Vilgalys and Hester 1990; White et al. 1990). These two regions were used as the main barcodes for taxonomic evaluation and identification. We also amplified the β-tubulin (tub2) and the second largest subunit of RNA polymerase II (rpb2) using the primer pairs Bt2a/Bt2b (Glass and Donaldson 1995) and RPB2-5F/RPB2-7R (Liu et al. 1999), respectively. The tub2 region was used to resolve the phylogeny of closely related species in Malbranchea and Gymnoascus, and the rpb2 region was only used to compare sister taxa in Malbranchea due to the limited availability of this region in public databases. Briefly, PCR conditions for ITS, LSU, tub2, and rpb2 consisted of an initial denaturation step at 95 °C for 5 min, followed by 35 cycles of 30 s at 95 °C, 1 min at 56 °C, and 1 min at 72 °C, with a final extension step at 72 °C for 10 min. PCR products were purified and sequenced at Macrogen Corp. Europe (Madrid, Spain) using the same primers employed for amplification. Consensus sequences were assembled using SeqMan v. 7.0.0 (DNAStar Lasergene, Madison, USA) and can be found in Table 1.

The resulting sequences were compared with those available at the National Center for Biotechnology Information (NCBI), using the Basic Local Alignment Search Tool (BLAST; https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed in January 2025). A maximum similarity level of > 98% with ≥ 90% sequence coverage was used to determine related GenBank sequences. Accession numbers were carefully examined to identify those corresponding to type strains of accepted species, and these were combined with data from the most recent phylogenetic studies to perform phylogenetic analyses (Kandemir et al. 2022; Torres-Garcia et al. 2023). The accession numbers of the onygenalean strains included in the phylogenetic analyses are listed in the Suppl. material 1. The sequence alignments generated in this study have been deposited in Zenodo (https://doi.org/10.5281/zenodo.15373205, accessed in May 2025).

Sequence alignments were made using the ClustalW algorithm (Thompson et al. 1994) in MEGA (Molecular Evolutionary Genetics Analysis) software v. 6.0 (Tamura et al. 2013) and refined with MUSCLE (Edgar 2004) or adjusted manually as needed. Each genomic DNA region was aligned individually before being combined into a single dataset. The phylogenetic concordance among individual trees was visually assessed to identify incongruent results among clades with high statistical support. Following confirmation of concordance, the separate alignments were concatenated into a single data matrix in MEGA (Tamura et al. 2013).

Maximum likelihood (ML) analyses were conducted using IQ-TREE v. 2.3.6 (Nguyen et al. 2015; Minh et al. 2020) with ultrafast bootstrapping for the estimation of branch support (Minh et al. 2013; Hoang et al. 2018). The most suitable evolutionary model for each partition was estimated using ModelFinder (Chernomor et al. 2016; Kalyaanamoorthy et al. 2017), implemented in IQ-TREE, following the Bayesian information criterion. A bootstrap value ≥ 95 was considered significant (Minh et al. 2013; Hoang et al. 2018). Bayesian inference (BI) analyses were run using MrBayes v. 3.2.6 (Ronquist et al. 2012). The best substitution model for each locus was estimated using jModelTest v. 2.1.3, following the Akaike criterion (Guindon and Gascuel 2003; Darriba et al. 2012). Markov chain Monte Carlo (MCMC) sampling was performed for 10 million generations using four simultaneous chains (one cold chain and three heated chains), starting from a random tree topology. Trees were sampled every 1000^th generation or until the run was stopped automatically when the average standard deviation of split frequencies fell below 0.01. The first 25% of the trees were discarded as the burn-in phase of each analysis, and the remaining trees were used to calculate posterior probabilities (pp), representing branch support for the consensus trees. A pp value ≥ 0.95 was considered significant (Hespanhol et al. 2019). The resulting trees were plotted using FigTree v. 1.3.1 (http://tree.bio.ed.ac.uk/software/figtree/, accessed in March 2025).

Malbranchea strains were characterized through species delimitation analyses. We performed a coalescent-based method, Bayesian Poisson Tree Processes, bPTP (Zhang et al. 2013), and a distance-based method, Assemble Species by Automatic Partitioning, ASAP (Puillandre et al. 2021). These analyses were performed on the concatenated ITS–LSU–tub2 alignment and the ITS single alignment. The input trees for the PTP analyses were inferred in IQ-TREE, as explained above. The analyses were conducted on the web server for bPTP (https://species.h-its.org/ptp/, accessed during November 2024) with 500,000 MCMC generations, thinning set to 100, burn-in of 10%, and a seed value of 123456. Convergence of the MCMC iterations was assessed by visualizing the log-likelihood trace plot, where the chain’s predominant exploration of high-likelihood locations was used as an indicator of convergence. The ASAP analyses were also performed for the concatenated and single ITS alignments using the matrix of pairwise nucleotide distances. This matrix was computed in MEGA v. 6.0 using the K2+G nucleotide substitution model. These analyses were conducted on the web server for ASAP (https://bioinfo.mnhn.fr/abi/public/asap/, accessed during November 2024). The resulting partition was determined by the lowest ASAP score. We also considered additional partitions when their ASAP score was identical to or close to the lowest value, as recommended (Puillandre et al. 2021). In such cases, we selected the closest partition to the delimitation predicted by phylogenetic analyses.

Hypothetical species delimitation was established by evaluating the phylogenetic analyses and the delimitation obtained through PTP and the ASAP partition. Therefore, the strains determined as unknown lineages were considered putative novel species and selected for further morphological studies.

The strains determined as putative novel species through phylogenetic analyses were selected for further morphological analyses. Additional strains from curated collections were requested based on their morphological resemblance and close phylogenetic relationship to the marine strains (Table 1). We examined six strains of Malbranchea spp. available in the CBS (Westerdijk Fungal Biodiversity Institute, Utrecht, The Netherlands) and BCCM/IHEM (Sciensano, Brussels, Belgium) collections, labeled as M. irregularis (CBS 149937), M. longispora (CBS 135817), M. multiseptata (CBS 146931), M. sinuata (CBS 149938), M. thaxteri (CBS 319.61), and Malbranchea sp. (IHEM 28255). The former four species have been recently published from freshwater sediments and other non-aquatic substrates (Rodríguez-Andrade et al. 2021; Torres-Garcia et al. 2023).

Macroscopic characterization of the colonies was conducted using various culture media, i.e., PDA, OA, and phytone yeast extract agar (PYE; Pronadisa, Madrid, Spain). Prepubertal human hair was deposited on the surface of OA plates at 25 °C to induce sexual morph production for those strains related to M. filamentosa and susceptible to heterothallism (Sigler et al. 2002). Color notations in descriptions followed the Methuen Handbook of Colour (Kornerup and Wanscher 1967). Microscopic characterization was made on OA after 14 days at 25 °C in the dark, unless otherwise specified in the description. Reproductive structures were mounted with lactic acid and observed under an Olympus BH-2 bright-field microscope (Olympus Corporation, Tokyo, Japan). Descriptions were based on a minimum of 30 measurements of the relevant structures to provide size ranges. Photomicrographs were taken using a Zeiss Axio-Imager M1 light microscope (Oberkochen, Germany) with Nomarski (interference contrast) and phase contrast condensers, coupled with a DeltaPix Infinity digital camera. Photoplates were assembled using Photoshop CS6 v.13.0. Scanning electron microscopy (SEM) micrographs were obtained using a Quanta 600 FEG Scanning Electron Microscope (Thermo Fisher Scientific, Waltham, USA), and specimens were processed according to the protocol established by Figueras and Guarro (1988).

The strains isolated in this study were preserved in the culture collection of the Faculty of Medicine in Reus (FMR, Reus, Catalonia, Spain), and those representing novel fungi were deposited at the Westerdijk Fungal Biodiversity Institute (CBS collection), including holotypes as dry colonies on the most appropriate media for their sporulation. Nomenclatural novelties were registered in MycoBank (https://www.mycobank.org/, accessed in November 2024).

Biogeographical distribution of the marine strains was assessed using the GlobalFungi database v.10.0, release version 5.0, accessed in December 2024 (Větrovský et al. 2020). This repository comprised data from 84,972 environmental samples originating from 846 studies (33% from Europe, 33% from Asia, 20% from North America, 4% from Australia, 2% from Africa, and 8% from other regions), encompassing terrestrial, airborne, aquatic, and marine samples. The database was organized with separate entries for the ITS1 and ITS2 regions, containing over 1 billion ITS1 sequences and more than 3 billion ITS2 sequences. First, we calculated the interspecific genetic distance of the ITS1 and ITS2 regions for each genus detected in marine sediments, following the methodology established for other fungal groups (Réblová et al. 2021b; Torres-Garcia et al. 2023). The resulting threshold, along with the full-coverage limit, was used to perform BLAST searches for each strain in GlobalFungi. Due to the large number of environmental sequences obtained, we selected representative sequences from a variety of geographical locations, biomes, and substrates. Sequence identifications were validated through phylogenetic analyses of the ITS1 and ITS2 regions separately. Metadata associated with each sample, including location, substrate, biome, mean annual temperature (MAT), mean annual precipitation (MAP), and pH, were used for ecological characterization. Geographic occurrences were represented using QGIS v.3.34.5 (QGIS Development Team 2024).

The ecological behavior of the onygenalean strains recovered from marine sediments was further tested in vitro, including the aforementioned strains from other collections. The physiological traits tested included salt tolerance and the ability to degrade cellulose, chitin, and keratin. Salt tolerance was assessed on MEA supplemented with 3.5, 10, 15, and 20% NaCl, using MEA as a negative control. The ability to degrade cellulose was evaluated on Cellulose Congo-Red Agar (CCA; 0.5 g KH₂PO₄, 0.25 g MgSO₄, 2 g cellulose, 0.2 g Congo-Red, 2 g gelatin, 15 g agar, 1 L distilled water, pH 6.8–7.2) (Gupta et al. 2012). Chitinase activity was assessed using Chitin Agar (CA; 20 g colloidal chitin, 1 g (NH₄)₂SO₄, 1 g K₂HPO₄, 0.5 g MgSO₄, 0.5 g KCl, 0.5 g NaCl, 0.2 g CaCl₂, 2 g FeSO₄.7H₂O, 15 g agar, 1 L distilled water) adjusted to pH 7 (Yilma and Kebede 2018). Colloidal chitin was obtained from commercial chitin (Sigma-Aldrich, Burlington, USA) through the modification of a published protocol (Hsu and Lockwood 1975). Briefly, 20 g of chitin was dissolved in 37% HCl by stirring at 40 °C, precipitated as a colloidal suspension by slowly adding 2 L of distilled water at 10 °C, collected by filtering through coarse filter paper, and washed with distilled water until pH 7. For these tests, colony diameter was measured after 21 days of incubation at 25 °C, and discoloration of Congo Red or CA was used as an indicator of positive cellulose- or chitin-degrading activity. Keratinophilic activity was tested by inoculating a suspension of fungal elements (ascospores, conidia, and hyphae) on a Petri dish containing 25 mL of distilled water, 1 mL of 10% yeast extract, and 1–1.5 cm of prepubertal child hair that had been previously washed with distilled water and sterilized (De Hoog et al. 2020). Plates were checked after 14 and 21 days, and the formation of perforating organs confirmed positive keratin degradation.

Finally, the ability of each putative novel species to grow at different temperatures was assessed on PDA from 5 to 45 °C at 5 °C intervals, including measurements at 37 °C, after 14 days of incubation. All physiological tests were performed in duplicate.

We recovered 32 strains from the primary cultures of sediment samples that were identified morphologically as members of the order Onygenales. Most of these strains were recovered through the direct plating of the sediment, with only twelve strains (37.5%) detected through the SMF pre-treatment. Regarding the culture media of isolation, 16 strains were isolated from PDA+C, nine from SWMEA3%, six from DRBC, and one from SWOA. Combining morphology and comparison of the ITS barcode region, one strain was recognized as a member of Myriodontium in Neoarthropsidaceae, four strains as members of the genera Aphanoascus and Byssoonygena in Onygenaceae, eight strains as members of the genus Malbranchea in Malbrancheaceae, and 17 strains as members of the genera Gymnascella, Gymnoascoideus, Gymnoascus, Narasimhella, and Sporendonema in Gymnoascaceae. Nevertheless, six strains of Malbranchea (FMR 19403, FMR 19564, FMR 20150, FMR 20151, FMR 20852, and FMR 21121) and one strain of Gymnoascoideus (FMR 19992) did not match any described species in their respective genera. The only marine strain that could not be identified at the genus/family level was FMR 20744, which exhibited a sexual morph resembling the genus Albidomyces, whereas the asexual morph showed similarities to the genera Amaurascopsis and Pseudoamaurascopsis (Guarro et al. 1992a; Torres-Garcia et al. 2023). All these strains represented 22 species under the consolidated species concept, combining phylogeny, morphology, and ecology as explained in the following sections.

The molecular identification and taxonomic position of the strains were assessed through phylogenetic analysis of the ITS and LSU regions. Each region was initially analyzed independently (Suppl. material 2: figs S1, S2), and after confirming the congruence between the topology of the resulting trees, both regions were concatenated into a single matrix. For the members of Gymnoascaceae and Malbrancheaceae, the final alignment encompassed 137 sequences representing ex-type and reference strains and also included members of Neogymnomycetaceae and a well-supported clade labeled as incertae sedis according to Kandemir et al. (2022). The outgroups selected were Polytolypa hystricis UAMH 7299 and Dactylodendron pinicola CBS 653.89. The alignment comprised 1,293 bp (700 bp for ITS and 593 bp for LSU), of which 584 positions were conserved (211 bp for ITS and 373 bp for LSU), and 671 were variable (465 bp for ITS and 206 bp for LSU). Among the variable sites, 544 were parsimony-informative (376 bp for ITS and 168 for LSU). The remaining positions included ambiguous or gap-only sites that were excluded from these categories. The models selected for the ML analysis were GTR+F+I+G4 for ITS and TIM3+F+R3 for LSU, while for the BI analysis, the best models were GTR+I+G for ITS and TIM1+I+G for LSU. The resulting phylogenetic tree confirmed the monophyly of the families included in the analyses with well-supported clades. Here, we represent the maximum-likelihood tree with the bootstrap support values of the ML analysis and the BI posterior probabilities at the nodes (Fig. 2). An additional phylogenetic tree was made to represent members of Onygenaceae and Neoarthropsidaceae, based on their great phylogenetic distance from the rest of the onygenalean strains recovered in this work (Suppl. material 2: fig. S3).

The strains recovered from the marine sediments were predominantly distributed in Gymnoascaceae, Malbrancheaceae, and Onygenaceae. The identification of each strain, along with its metadata, can be accessed in Table 1. In Onygenaceae, the following species were detected: Aphanoascus (A.) crassitunicatus (FMR 20176), A. fulvescens (FMR 20116, FMR 20522), and Byssoonygena (B.) ceratinophila (FMR 19558); while a single species was detected in Neoarthropsidaceae, Myriodontium (My.) keratinophilum (FMR 20839) (Suppl. material 2: fig. S3).

The family Gymnoascaceae was the most represented (Fig. 2), with the following species identified: Gymnascella (Gl.) dankaliensis (FMR 19616, FMR 19625, and FMR 20162), Gl. nodulosa (FMR 20375), Gl. udagawae (FMR 20467, FMR 20506, and FMR 20769); Gymnoascus (G.) reessii (FMR 19395); Narasimhella (N.) echinulata (FMR 20851), N. hyalinospora (FMR 20283), N. poonensis (FMR 19613, FMR 19620, FMR 19621, and FMR 19978); and Sporendonema (S.) casei (FMR 21361). However, the analysis of the ITS and LSU regions was inconclusive for the precise identification of the Gymnoascus strains FMR 20452 and FMR 21364. Therefore, we performed an additional concatenated ITS–LSU–tub2 analysis of the genus, which confirmed these strains as G. exasperatus and G. uncinatus, respectively (Suppl. material 2: fig. S4).

In Gymnoascaceae, this ITS–LSU analysis resolved the strain FMR 19992 as an independent lineage in Gymnoascoideus (Gd.), phylogenetically related to the type strain of Gd. boliviensis (IMI 351986; 91% identity with ITS, 97% identity with LSU) and Gd. petalosporus (UAMH 1665; 91% identity with ITS, 97% identity with LSU). At the same time, the unidentified strain FMR 20744, together with the epitype strain of Diploospora (Di.) rosea (DAOM 250100), formed a well-supported incertae sedis clade in Onygenales. However, the former strain delineated an independent lineage with sufficient phylogenetic distance from Di. rosea (76% identity with ITS and 92% identity with LSU) to be considered a distinct taxon. Therefore, strain FMR 19992 is proposed as Gd. alboluteus sp. nov., and FMR 20744 is proposed as the novel genus Deilomyces (D.), typified with D. minimus gen. et sp. nov. The detailed morphological characterizations of these fungi are provided in the taxonomy section.

Among the Malbranchea (M.) marine strains, only FMR 20086 and FMR 20173 could be confirmed as M. zuffiana and M. ostraviensis, respectively. Since the ITS/LSU analysis was unable to confidently assign the remaining six strains to any known species, we carried out additional phylogenetic analyses for species delineation. The unidentified strains were labeled as Malbranchea sp. 1–4 in Fig. 2.

The concatenated ITS–LSU–tub2 analysis resolved the phylogeny of the genus Malbranchea and delineated the relationships of the unidentified strains (Fig. 3). After confirming the congruence of the three regions, the final alignment comprised 1597 bp (631 bp for ITS, 581 bp for LSU, and 385 bp for tub2), of which 929 positions were conserved (312 bp for ITS, 447 bp for LSU, and 170 bp for tub2), and 635 were variable (302 bp for ITS, 127 bp for LSU, and 206 bp for tub2). Among the variable sites, 468 were parsimony-informative (216 bp for ITS, 101 bp for LSU, and 151 bp for tub2). The remaining positions included ambiguous or gap-only sites that were excluded from these categories. The models selected for the ML analysis were SYM+G4 for ITS, TNe+R2 for LSU, and TNe+I+G4 for tub2, while for the BI analysis, the best models were SYM+G for ITS, TIM1ef+I+G for LSU, and HKY+I+G for tub2. The outgroups selected were Auxarthronopsis bandhavgarhensis CBS 134524 and Neogymnomyces demonbreunii CBS 427.70. The representative strain of M. cinnamomea, DTO 403-H2, was excluded from the analysis based on its great phylogenetic distance from other Malbranchea species.

The resulting phylogenetic tree determined four major, highly supported clades in Malbranchea, where our strains were distributed across clades I, II, and III (Fig. 3). Each strain was compared with its closest phylogenetic taxa for the percentage of identity of the ITS, LSU, tub2, and rpb2 regions, when available. In clade I, strain FMR 20151 was grouped with the unidentified Malbranchea strain (IHEM 28255), isolated from bat fur. These two strains represented an independent lineage phylogenetically related to the type strain of M. thaxteri (CBS 248.58; 97% identity with ITS, 93% identity with tub2, and 94% with rpb2) and the type strains of M. filamentosa (UAMH 4097; 97% identity with ITS) and M. stricta (CBS 146932; 97% identity with ITS, 99% identity with LSU). In the same clade, strain FMR 20852 formed a singleton lineage related to the type strains of M. gymnoascoides (CBS 146930; 97% identity with ITS, 99% identity with LSU), M. ostraviensis (CCF 4241; 97% identity with ITS, 96% identity with tub2), M. phuphaphetensis (CV00115; 92% identity with ITS, 99% identity with LSU, 96% identity with tub2), and M. umbrina (UAMH 3952; 94% identity with ITS, 99% identity with LSU, 96% identity with tub2). In clade II, the strains FMR 19564, FMR 20150, and FMR 21121 were grouped with strain CBS 319.61, which was labeled as M. thaxteri. These strains formed a well-supported clade related to the type strains of M. sinuata (CBS 149938; 98% identity with ITS, 99% identity with LSU, 96% identity with tub2, and 95% with rpb2) and M. albolutea (CBS 125.77; 96% identity with ITS, 99% identity with LSU, 93% identity with tub2), but were placed distantly from the type strain of M. thaxteri included in clade I. In clade III, strain FMR 19403 formed a singleton, distant lineage related to the type strains of M. concentrica (CBS 112861; 92% identity with ITS, 94% identity with tub2), M. longispora (CBS 135817; 95% identity with ITS, 99% identity with LSU, 96% identity with tub2), and M. multiseptata (CBS 146931; 94% identity with ITS, 99% identity with LSU). These marine lineages appeared to represent four putative novel species based on their substantial phylogenetic distances from their respective closest species. This hypothesis was further tested through speciation analyses.

The ASAP and bPTP speciation analyses were performed on both the concatenated and single ITS alignments of Malbranchea (Suppl. material 2: fig. S5). Both the concatenated and ITS ASAP tests determined each marine lineage as an independent partition, and the results for the other Malbranchea species were congruent with the phylogenetic analysis. For the bPTP tests, the resulting topology for the concatenated alignment also recognized each marine lineage as independent. However, the posterior probabilities for speciation were low in clades expected to represent speciation events, suggesting that our data might not conform well to the software’s algorithm.

Therefore, based on phylogenetic evidence, morphological features, and the ecological data presented below, we concluded that Malbranchea sp. 1–4 represented novel species within the genus. These are proposed as M. parafilamentosa sp. nov. (FMR 20151 and IHEM 28255; Malbranchea sp. 3 in Fig. 2), M. sedimenticola sp. nov. (FMR 19564, FMR 20150, FMR 21121, and CBS 319.61; Malbranchea sp. 1 in Fig. 2), M. seminuda sp. nov. (FMR 19403; Malbranchea sp. 2 in Fig. 2), and M. sexualis sp. nov. (FMR 20852; Malbranchea sp. 4 in Fig. 2). A detailed morphological characterization of these fungi is provided in the taxonomy section.

The majority of the onygenalean species isolated in this study were recovered from Arrabassada and Miracle beaches and from the discharge area of the Ebro River (Table 1).

At Arrabassada Beach, nine species were identified, with A. crassitunicatus, Gd. alboluteus, M. ostraviensis, M. parafilamentosa, and N. hyalinospora exclusively detected at this location. Four species were recovered only from Miracle Beach, i.e., B. ceratinophila, G. reessii, M. seminuda, and M. zuffiana. Three additional species—Gl. dankaliensis, N. poonensis, and the novel species M. sedimenticola—were detected at both Arrabassada and Miracle beaches. Aphanoascus fulvescens was found at Arrabassada Beach and in the discharge area of the Ebro River. The sediments collected in the Ebro River area revealed a distinct community of onygenalean fungi, primarily dominated by members of Gymnoascaceae. This group was represented by G. exasperatus, N. echinulata, and two species of Gymnascella—Gl. nodulosa and Gl. udagawae—along with the novel species M. sexualis. In the discharge area of the Llobregat River, we detected G. uncinatus, M. sedimenticola, and S. casei. In contrast, only two species were identified in the Ter River discharge area: My. keratinophilum and the novel taxon D. minimus.

These species were predominantly (77%) recovered from sediments collected at depths between 27 and 30 m. Notably, N. poonensis, Gl. dankaliensis, and Gl. udagawae—the most prevalent species in our study (n = 4, n = 3, and n = 3, respectively)—were almost exclusively detected at 27 m depth. Although the novel species were less abundant, they were also primarily recovered from the deepest sediments. The only exceptions were M. sedimenticola, also recovered from sediments at 19 m depth and from soil (CBS 319.61), and M. parafilamentosa, also isolated from bat fur (IHEM 28255). Based on this substrate variability, we assessed geographic and substrate distribution using the GlobalFungi database.

Prior to biogeographic analysis, environmental sequences were clustered with curated sequences in individual alignments. Environmental sequences were obtained from GlobalFungi using a threshold we determined at 98% interspecific distance for the ITS1 and ITS2 regions of Aphanoascus, Byssoonygena, Gymnoascoideus, Gymnoascus, Gymnascella, Malbranchea, Myriodontium, Narasimhella, and Sporendonema. For the novel genus Deilomyces, we used a 95% identity threshold, based on its great distance to its sister lineage (Diploospora) and the poor representation of these taxa in public databases. BLAST searches on GlobalFungi were conducted according to these thresholds, and the selected environmental sequences were included in the alignments to ensure phylogenetic identification (Suppl. material 2: figs S6–S17). The species with the highest number of environmental sequences were A. fulvescens (Suppl. material 2: fig. S6), Gl. dankaliensis (Suppl. material 2: fig. S9), G. reessii (Suppl. material 2: figs S11, S12), and My. keratinophilum (Suppl. material 2: fig. S15). Additionally, the phylogenetic analysis of the ITS2 region of Gymnoascoideus revealed three environmental sequences that clustered as an independent hidden taxon, labeled ITS2-ENV1 (Suppl. material 2: fig. S10). Unfortunately, these sequences were not detected in the ITS1 region, where only a single sequence was related to strain FMR 19992 (Suppl. material 2: fig. S10). For the genus Narasimhella, only the ITS1 region was used in the biogeography analysis (Suppl. material 2: fig. S18). This region detected environmental sequences exclusively associated with N. poonensis. Other species in the genus—N. armeniaca, N. echinulata, N. hyalinospora, and N. punctata—showed close phylogenetic relationships, making confident assignment of environmental sequences impossible. Analysis of the ITS2 region revealed a similar pattern (identity ≥ 98% between species), so ITS2 was excluded from the biogeographic analysis.

The search on GlobalFungi revealed that some species appeared to represent common fungi with a worldwide distribution, while others were rare species with a more restricted biogeography. Most species of Malbranchea, as well as A. fulvescens, D. minimus, G. reessii, Gd. alboluteus, and My. keratinophilum, were detected in both hemispheres (Fig. 4A; Suppl. material 2: fig. S18). Deilomyces minimus was found to be a frequent fungus (more than 1,500 observations), particularly in tropical areas of America, Africa, and Asia (Fig. 4A). For Malbranchea, M. ostraviensis and M. zuffiana were the most common species across different environmental samples, with M. zuffiana clustering in South Africa and eastern China (Fig. 4A). The novel species M. sexualis and M. sedimenticola showed a narrower range of distribution, and M. seminuda was detected as a rare fungus (fewer than 150 observations), with a geographic distribution restricted to the area between 30°–40° North and South (Fig. 4A). A similar pattern of restricted geographical distribution was also observed in Gymnoascoideus. The novel species Gd. alboluteus was also a rare fungus (fewer than 50 observations), detected only in Europe and Australia (Fig. 4A). The novel species M. parafilamentosa, recovered from marine sediments and bat fur (Table 1), represented another rare fungus that, in this case, could not be associated with any environmental sequence in GlobalFungi.

The species detected in GlobalFungi were predominantly terrestrial, primarily associated with soil-related substrates such as bulk soil, rhizosphere soil, topsoil, and plant roots (Fig. 4B). The species that exhibited wide geographic distribution appeared to endure a variety of environmental conditions. In particular, D. minimus, M. ostraviensis, and M. zuffiana were predominantly detected in soil-related samples from croplands and woodlands of temperate to tropical regions (Fig. 4A, B). These areas were characterized by mean annual temperatures (MATs) between 4–20 °C and mean annual precipitations (MAPs) between 200–1,000 mm, including North America, the Caribbean, southern Europe, central and southern Africa, eastern China, and Australia. Malbranchea sexualis was also predominantly associated with soil samples from temperate and tropical areas in Europe, North America, China, and Australia (Fig. 4A, B), but it endured colder MATs between -5 °C and 20 °C. Among the species with more restricted geographic distribution, M. seminuda was exclusively detected in soil samples from grasslands with MATs around 15 °C and MAPs around 700 mm in southern Europe, the western United States, and southern Africa (Fig. 4A, B). Malbranchea sedimenticola was associated with soil-related substrates from deserts and shrublands with MATs between 7 °C and 20 °C and MAPs between 20 mm and 500 mm, on the west coast of the American continent, southern Europe, western Australia, and northern Africa (Fig. 4A, B). For the species of Aphanoascus, Byssoonygena, Gymnascella, Gymnoascus, Myriodontium, and Sporendonema, the geographical distribution observed in GlobalFungi predominantly covered terrestrial areas and was also associated with soil-related substrates. Among them, A. fulvescens and G. reessii were the most abundant and prevalent species across terrestrial samples (Suppl. material 2: figs S18, S19).

Nevertheless, several species were also detected in water-related substrates (Fig. 4B; Suppl. material 2: fig. S19). The species Gl. nodulosa, M. sexualis, and S. casei were strongly associated with aquatic plant shoots. Freshwater sediment was another common water-related substrate, with A. crassitunicatus, Gl. dankaliensis, Gl. udagawae, G. reessii, and N. poonensis detected in China (Suppl. material 2: figs S18, S19), and D. minimus in Australia (Figs 4A, B). Finally, marine sediments represented an unusual substrate, but the species A. fulvescens, Gl. dankaliensis, and My. keratinophilum were detected in the Pacific Ocean, and G. reessii in the Atlantic Ocean (Suppl. material 2: figs S18, S19).

Based on the dominant terrestrial prevalence detected through GlobalFungi for the species recovered from marine sediments, we aimed to test in vitro their ability to grow under different conditions resembling both terrestrial and marine environments. The results of the physiological tests are represented in Table 2 and Fig. 5. Most strains grew in media supplemented with up to 10% NaCl. Only the strains FMR 19625 (Gl. dankaliensis), FMR 19564 (M. sedimenticola), CBS 149938 (M. sinuata), and FMR 20151 (M. parafilamentosa) tolerated up to 15% NaCl, and My. keratinophilum (FMR 20839) was the only species that grew up to 20% NaCl. In contrast, the strains of D. minimus (FMR 20744), M. seminuda (FMR 19403), and M. multiseptata (CBS 146931) did not grow in any media supplemented with NaCl. Regarding the test on cellulose degradation, all strains grew on CCA and were able to degrade cellulose, producing a halo around the colony. For the test assessing chitin degradation, all strains managed to grow on CA, except D. minimus. However, degradation halos were only observed in the strains of M. parafilamentosa (FMR 20151 and IHEM 28255), M. sexualis (FMR 20852), M. sinuata (FMR 18266), My. keratinophilum (FMR 20839), and S. casei (FMR 21361). Finally, all strains grew in the keratin test, but hair degradation was only observed in the strains of A. crassitunicatus (FMR 20176), A. fulvescens (FMR 20116, FMR 20522), and B. ceratinophila (FMR 19558). The production of perforating organs can be observed in Fig. 5C, D.

The geographical distribution of fungi has been a long-standing question in mycology. Advances in DNA-based methods have significantly enhanced our understanding of species boundaries, increasing fungal diversity and challenging the traditional notion that microscopic fungi are ubiquitous (Hawksworth and Lücking 2017; Niskanen et al. 2023). Metabarcoding analyses have determined Onygenales as a small but significant component of soils worldwide, with Onygenaceae and Gymnoascaceae as the most prevalent families (Coleine et al. 2022). Our search in GlobalFungi supports these observations, with all species recovered from marine sediments displaying a strong association with soil and terrestrial habitats. It is noteworthy that, in a similar fashion to soil, Gymnoascaceae was the most prevalent and diverse family in marine sediments. Members of Onygenaceae, Neoarthropsidaceae, and Malbrancheaceae were also detected, with the latter as a particularly diverse group (Table 1). Among this diversity, some species displayed a ubiquitous distribution. The species A. fulvescens, G. reessii, and My. keratinophilum represented abundant and widespread fungi in both hemispheres (Suppl. material 2: fig. S18); their frequent detection in culture-dependent surveys emphasizes their high colonization capacity and environmental prevalence (Guarro et al. 2012). The geographical distribution of the remaining species of Aphanoascus and Gymnoascus detected in marine sediments covered both hemispheres, but their abundance was notably lower. In contrast, the species of Malbranchea and Gymnoascoideus displayed restricted patterns of distribution.

The genus Malbranchea represents one of the most prevalent groups of Onygenales in soil (Coleine et al. 2022). The species M. ostraviensis and M. zuffiana were detected around the globe. It is surprising that these species can be detected abundantly in environmental samples through metabarcoding but are rarely recovered in culture. Indeed, M. ostraviensis was described from clinical samples before evidencing its presence in the environment (Hubka et al. 2013). In contrast, other species displayed a restricted distribution, such as M. seminuda in areas with Mediterranean climates and M. sedimenticola in desertic areas and shrublands (Fig. 4) (Olson et al. 2001). This latter species was able to grow up to 37 °C, supporting this limited range of distribution. In addition, it represents a sister lineage to M. sinuata, a species that, despite its original description as a singleton from Spanish freshwater sediments, has been exclusively detected in soil samples from Spain through GlobalFungi (Torres-Garcia et al. 2023). In fact, there are species of Malbranchea that are not currently represented in this database, such as M. parafilamentosa, M. echinulata, and M. irregularis, the last two only known from Spanish freshwater sediments (Torres-Garcia et al. 2023). The absence of these species in GlobalFungi suggests their low abundance in nature and their preference for uncommon habitats, including aquatic and animal-associated substrates.

The novel species Gymnoascoideus alboluteus represents another uncommon fungal taxon. It was detected in GlobalFungi at low abundance from three forest soil samples originating from Europe, China, and Australia. The considerable geographic distance among these sites suggests a wide distribution range; however, its low abundance may hinder consistent detection. Other species within the genus, such as Gd. petalosporus and Gd. boliviensis, have been sporadically detected from soil, as well as from human clinical specimens and dung (Orr et al. 1977; Guarro et al. 1992b). These observations support the notion that Gymnoascoideus comprises a group of rare environmental fungi.

Conversely, we also identified the opposite scenario: fungi that are abundant in the environment but have never been recovered in culture. Analysis of the ITS2 in GlobalFungi revealed a new lineage, ITS2-ENV1, that is sister to Gd. alboluteus (Suppl. material 2: fig. S10) and associated with soil habitats. This lineage is currently known only from sequence data. Such taxa are classified as “dark matter fungi” due to the absence of voucher specimens or cultured isolates, which hampers their formal description (Lücking et al. 2021). The number of dark matter fungi uncovered through metabarcoding is rapidly increasing, driven by both the detection of fungi that are difficult to culture and the exploration of understudied environments, such as aquatic ecosystems (Grossart et al. 2016; Torres-Garcia et al. 2022).

These findings underscore the importance of integrating metabarcoding and culture-based approaches in fungal biodiversity studies. Overall, the GlobalFungi database is a valuable resource for assessing fungal distribution using high-throughput sequencing metabarcoding data. However, it revealed two main limitations: (1) a geographical bias toward samples from the Northern Hemisphere and (2) limited species-level resolution for certain taxa when using only the ITS1 or ITS2 regions (Réblová et al. 2021b). Our exploration of GlobalFungi displayed both widespread taxa and species restricted to specific environmental conditions (Fig. 4; Suppl. material 2: fig. S18). Moreover, ITS1 and ITS2 phylogenetic trees revealed that each region alone was sufficient to differentiate onygenalean species, consistent with prior studies (Moulíková et al. 2023; Torres-Garcia et al. 2023). The only exception, in our case, was the ITS2 region of Narasimhella. The species-level resolution of the ITS barcode varies across fungal lineages: while ITS is insufficient for differentiating members of Chaetomiaceae and Podosporaceae, it remains the primary barcode for identifying Codinaea, Zanclospora, and most Onygenales (Wang et al. 2019, 2022; Réblová et al. 2021a, 2021b; Kandemir et al. 2022; Torres-Garcia et al. 2023). Thus, we conclude that the limitations of GlobalFungi were effectively addressed in our study, and onygenalean fungi are suitable for biogeographical analysis using the ITS region.

The search in GlobalFungi confirmed the marine environment as an underexplored habitat. The species A. fulvescens, G. reessii, Gl. dankaliensis, and My. keratinophilum were associated with marine substrates, particularly in shallow coastal areas (Suppl. material 2: figs S18, S19). Previous surveys in shallow marine sediments from the coasts of Catalonia (western Mediterranean Sea) explored the efficacy of cycloheximide and the hair-baiting method to recover onygenalean fungi in culture (Ulfig et al. 1997, 1998). The hair-baiting method selected fungi that degrade keratin, such as A. fulvescens and A. keratinophilus (Ulfig et al. 1997). On the other hand, the use of cycloheximide in culture media revealed a broader diversity of onygenalean fungi, with species such as A. fulvescens, Gl. dankaliensis, Gl. littoralis, M. pulchella, M. conjugata, M. reticulata, M. umbrina, M. zuffiana, My. keratinophilum, N. hyalinospora, and N. poonensis (Ulfig et al. 1998). In this work, we detected most of these species using two isolation methods and four different culture media. Interestingly, most of the species were recovered in culture from sediments at 27 m of depth through the direct plating of the sediment and using cycloheximide (PDA+C) (Table 1). The addition of cycloheximide played a crucial role in unraveling the diversity of onygenalean fungi in marine sediments. This protein-synthesis inhibitor prevented the growth of most environmental fungi, facilitating the detection of these ascomycetes (Wingfield et al. 2022). In fact, most of the strains representing novel species were isolated with PDA+C, independently of the isolation method, direct plating or SMF (Table 1). However, SMF enabled the detection of singleton rare species such as D. minimus and M. seminuda. The medium PDA+C has also been used to recover Onygenales from freshwater sediments of the rivers Ter and Llobregat (Torres-Garcia et al. 2023). These rivers were characterized by species of Malbranchea and Myriodontium, with M. ostraviensis and My. keratinophilum as the only species that we also detected in marine sediments. The detection of a different community while using the same isolation medium suggests an underlying process of adaptation to different environmental conditions.

Regarding the other culture media used for isolation, we detected A. fulvescens at 27 m of depth using SWMEA3% and demonstrated its ability to colonize human hair (Tables 1, 2). In addition, we detected an additional keratinophilous species, B. ceratinophila, at 13 m of depth using DRBC. The detection of this species was surprising because it had not been recovered in culture since its original description from garden soil (Guarro et al. 1987).

Our results emphasize the importance of employing various culture media in diversity surveys. Although onygenalean fungi tend to specialize in keratin degradation, many lineages display different nutritional preferences (Muñoz et al. 2018; Kandemir et al. 2022). In this way, cycloheximide outperforms the hair-baiting method in detecting Onygenales diversity in marine sediments. In this sense, seawater-based media (SWMEA3% and SWOA) exhibited a similar effect to PDA+C, facilitating the detection of halotolerant species over Other fungi.

The Onygenales can be divided into two main groups based on nutrition preferences: those degrading cellulose, which is considered an ancestral trait, and those that have specialized in keratin degradation (Kandemir et al. 2022). During Onygenales evolution, cellulases were reduced in favor of keratinases, enabling the colonization of animal-associated substrates (Muñoz et al. 2018). However, some onygenalean fungi can still produce both groups of enzymes. For example, M. zuffiana can be recovered as an environmental fungus but is also associated with animal and human clinical specimens (Rodríguez-Andrade et al. 2021). This species is known to predominantly produce keratinases but still maintains cellulolytic domains encoded in its genome (Granados-Casas et al. 2024). Other species that we have detected in marine sediments, like Gl. dankaliensis and M. ostraviensis, also represent common environmental fungi and causal agents of onychomycosis in humans (Fig. 4, Suppl. material 2: figs S18, S19) (Hubka et al. 2013; de Hoog et al. 2020; Rodríguez-Andrade et al. 2021). In the current study, these species only displayed cellulolytic activity in culture (Table 2), which might be related to the environmental origin of these strains. However, cellulolytic activity and hair degradation were only observed in A. crassitunicatus, A. fulvescens, and B. ceratinophila, all members of Onygenaceae (Table 2). This family also includes Coccidioides, one of the most relevant genera of dimorphic fungi responsible for systemic infections in humans (de Hoog et al. 2020; Kandemir et al. 2022). The ability of members of this family to degrade both cellulose and keratin, along with their presence in the environment, supports previous findings in clinical mycology suggesting that the environment may serve as a reservoir for clinically relevant fungi (Luberto et al. 2024).

Aquatic environments are characterized by the dominance of chitin as the main polysaccharide (Souza et al. 2011). Chitin is massively produced in the water column but barely detected in sediments, suggesting its degradation (Poulicek and Jeuniaux 1988; Souza et al. 2011). At the same time, aquatic plants and macroalgae represent a source of cellulose (Singhania et al. 2025). It is noteworthy that cellulose and chitin share similar chemical structures; therefore, chitin degradation may follow a similar pathway to cellulose degradation (Wörner and Pester 2019). All the species detected here degraded cellulose, and, in addition, we detected four species with chitinolytic activity from marine sediments: M. parafilamentosa, M. sexualis, My. keratinophilum, and S. casei; and one species from freshwater sediments, M. sinuata (Table 2). Although fungi are known to produce chitinases for maintaining their cell wall, the annotation of the genome of M. zuffiana revealed a great abundance of the GH18 enzyme family, which includes several exogenous chitinases (Granados-Casas et al. 2024; Paudel et al. 2025). The active degradation of chitin by bacteria has been demonstrated in both freshwater and marine sediments (Wörner and Pester 2019). Therefore, onygenalean fungi could play a similar role in these environments.

Salinity represents a physiological barrier to the colonization of the marine environment. Halotolerance has been a major driver of evolution across the eukaryotic tree of life, with multiple transitions between marine and non-marine lineages (Jamy et al. 2022). In fungi, halotolerance represents a common trait among terrestrial lineages, particularly within families like Gymnoascaceae (Gunde-Cimerman et al. 2009; Kandemir et al. 2022). This family was the most abundant and diverse in marine sediments (Table 1). Species of Gymnoascus and Narasimhella grew equally well in simple media and media supplemented with 3.5% salt, but species of Gymnascella grew better in media supplemented with 3.5% salt. Most of the members of Malbrancheaceae and Onygenaceae also tolerated up to 10% NaCl, with the exception of D. minimus and M. seminuda, which evaded growing on salty media (Table 2). This suggests that those species that tolerated salt could represent metabolically active fungi. In fact, there is an onygenalean species that has been exclusively recovered from the marine environment, Gl. littoralis (Landy and Jones 2006; Jones et al. 2019; Kandemir et al. 2022). Although the physiology of marine Onygenales remains obscure, it is noteworthy that, in this work, most of the species that degraded chitin, i.e., M. sinuata, M. parafilamentosa, M. sexualis, My. keratinophilum, and S. casei, also tolerated salt concentrations above marine-like conditions (Table 2).

The diversity of Onygenales that we recovered from marine sediments is composed of fungi predominantly associated with terrestrial habitats (Fig. 4, Suppl. material 2: figs S18, S19). Fungal propagules are known to remain viable for extended periods and can be dispersed through the air from terrestrial to marine environments, where they seem to accumulate in sediments (Fröhlich-Nowoisky et al. 2012). Despite the terrestrial origin of this community, the observed halotolerance in most species suggests the ability to adapt metabolically to marine conditions. For example, fungal spores from the Dead Sea have been reported to tolerate higher osmotic pressure than their terrestrial counterparts, suggesting a selection for osmotolerant propagules (Kis-Papo et al. 2003). Structural cell wall rearrangements have been reported in fungi recovered from the marine environment, enabling adaptation to both hypersaline and salt-deprived conditions and providing a mechanism for dealing with external stress (Fernando et al. 2023). Transition events between terrestrial and freshwater habitats to the marine environment seem to have occurred more frequently within fungi than in any other microbial eukaryotes (Jamy et al. 2022; Cunliffe 2023). Thus, the community detected here could represent fungi with the metabolic capacity to survive in both terrestrial and aquatic environments.

The ecological interpretation of the species detected in marine sediments is dependent on the taxonomic analyses developed for identification. For this reason, we identified the strains at the species level, combining morphology and phylogenetic analyses. The order Onygenales has primarily been studied using the ITS region since the advent of molecular methods (Untereiner et al. 2004; Arunmozhi Balajee et al. 2007). While this region has been effective in distinguishing species, many group boundaries remained unresolved. In response, the mycological community has made efforts to incorporate protein-coding regions such as tub2, rpb2, tef1, and actin into phylogenetic analyses (Kandemir et al. 2022). Unfortunately, there are few public sequences available for some of these genetic markers. In this study, we examined the relationships of marine sediment strains using both ITS and LSU regions (Fig. 2). In addition, we used the tub2 marker to clarify the phylogenies of Gymnoascus (Suppl. material 2: fig. S4) and Malbranchea (Fig. 3), strengthening the backbone support of the phylogenies and resolving relationships among sister taxa. There are rpb2 sequences available for several species of Malbranchea. However, their low number hindered a concatenated phylogenetic analysis, and we could only use them to compare percentages of identity.

The morphological and phylogenetic analyses of Malbranchea revealed the lack of robust morphological features between species, making morphology-based identifications challenging. Species such as M. albolutea and M. sedimenticola, which represented sister taxa (Clade II, Fig. 3), exhibited the characteristic features of the genus, including an Auxarthron-like sexual morph with peridial appendages and a sinuous arthroconidial asexual morph. These species were also nearly identical to M. thaxteri, a phylogenetically distant species (Clade I, Fig. 3). Meanwhile, species like M. seminuda, M. concentrica, and M. kuehnii (Clade III, Fig. 3) displayed different ascomata features, consisting of subtle peridial hyphae lacking appendages—a morphology that resembles the genus Amauroascus rather than Malbranchea (Solé et al. 2002). Nevertheless, this clade also included species with Auxarthron-like sexual morphs, such as M. longispora. Despite these morphological variations, our phylogenetic analysis (Figs 3, 4), as well as those from other recent studies, supported Malbranchea as a monophyletic genus (Kandemir et al. 2022; Torres-Garcia et al. 2023). Given these findings, we attempted to assess speciation events within Malbranchea, but unfortunately, posterior probabilities indicated that our data were insufficient for these analyses (Suppl. material 2: fig. S5). Since the recent re-erection of Malbranchea, there has been a lot of attention on this genus, and up to seven novel species have been described from freshwater sediments, cave soil, and clinical samples (Rodríguez-Andrade et al. 2021; Preedanon et al. 2023; Torres-Garcia et al. 2023). In this work, we proposed four additional species (Table 2). This increase in diversity suggests that the exploration of uncommon environments, in particular aquatic habitats, may reveal more Malbranchea species. Future descriptions of novel taxa in this genus should include protein-coding regions in the phylogenies, at least the tub2 region, to provide phylogenetic support to the different clades.

The genus Gymnoascoideus represents a poorly represented group in phylogenetic analyses of Onygenales (Kandemir et al. 2022). Although the type species, Gd. petalosporus, has been abundantly recovered in culture, only a couple of strains have been sequenced (Fig. 2). Since the description of Gymnoascoideus, only one additional species has been added to the genus, Gd. boliviensis (Kandemir et al. 2022). In this work, we introduced a third species, Gd. alboluteus, which seems to represent a rare environmental fungus (Fig. 4).

We also introduced the novel genus Deilomyces. According to our phylogenetic analysis, D. minimus and Di. rosea represented two distinct and distant genera placed in a well-supported lineage that remains unresolved at the family level (Fig. 2) (Kandemir et al. 2022). Diploospora rosea represents an ecologically rare fungus associated with cardboard and parchments that has only been recovered three times in culture (Tanney et al. 2015). In addition, it displays a mixture of thallic and blastic conidiogenesis with conidia arising in acropetal chains—a reproductive strategy distinct from the Chrysosporium-like asexual morph observed in Deilomyces. The great diversity within this small clade highlights the significant gaps in our understanding of certain onygenalean groups and emphasizes the importance of ongoing surveys to explore fungal diversity in environmental substrates.