IMA genome-F18

and Penicillium solitum , isolated from Italian salami


Introduction
Sequencing fungal genomes has now become very common and the list of genomes in this manuscript reflects this.Particularly relevant is that the first announcement is a re-identification of Penicillium genomes available on NCBI.The fact that more than 100 of these genomes have been deposited without the correct species names speak volumes to the fact that we must continue training fungal taxonomists and the importance of the International Mycological Association (after which this journal is named).When we started the genome series in 2013, one of the essential aspects was the need to have a phylogenetic tree as part of the manuscript.This came about as the result of a discussion with colleagues in NCBI who were trying to deal with the very many incorrectly identified bacterial genomes (at the time) which had been submitted to NCBI.We are now in the same position with fungal genomes.Sequencing a fungal genome is all too easy but providing a correct species name and ensuring that the fungus has in fact been correctly identified seems to be more difficult.We know that there are thousands of fungi which have not yet been described.The availability of sequence data has made identification of fungi easier but also serves to highlight the need to have a fungal taxonomist in the project to make sure that mistakes are not made.Introduction Penicillium and its 536 accepted species represent one of the most commonly occurring and important fungal genera (Houbraken et al. 2020;Visagie et al. 2014).In recent years, whole genome sequencing efforts have increased and hundreds of Penicillium genomes are publicly available in the NCBI genome database (https:// www.ncbi.nlm.nih.gov/ datas ets/ genome).The study of these genomes is important, for example, to gain a better understanding of the biology of certain species.However, these studies and their communication depend on the use of the correct name of the genomes and the conclusions drawn from them.Analyses such as genome comparisons based on incorrect identifications lead to incorrect conclusions.The problem of misidentified genomes has already been highlighted by Houbraken et al. (2021), who also made several recommendations to prevent misidentifications in future.To support future studies using the genomes currently available in NCBI with the name Penicillium, we re-identify the genomes here using the modern taxonomy of the genus as published in Houbraken et al. (2020) who published an accepted species list and an updated subgeneric classification at the subgenus, section and series levels.

Materials and methods
A Penicillium reference dataset was compiled mainly based on the most recent taxonomy and accepted species list published by Houbraken et al. (2020).The six gene regions included in the dataset were beta-tubulin (BenA), calmodulin (CaM), RNA polymerase II second largest subunit (RPB2), RNA polymerase II largest subunit (RPB1), the subunit of the cytosolic chaperonin Cct ring complex (Cct8), and Tsr1, the protein required for processing 20S pre-rRNA in the cytoplasm.These gene regions were extracted from genomes downloaded for Penicillium from the NCBI Genome Portal using Geneious Prime v. 2023.1.2and included in the dataset.
In our multi-gene phylogenetic analysis, each gene region was treated as separate partitions and introns and exons were taken into consideration where appropriate.Datasets were aligned using MAFFT v. 7.490 with the G-INS-i option (Katoh and Standley 2013).Alignments were trimmed or adjusted as needed and then concatenated in Geneious Prime.The General Time Reversible nucleotide substitution model with gamma distribution with invariant site (GTR + G + I) was chosen for all partitions.Maximum likelihood trees were calculated in IQtree v. 2.1.3(Minh et al. 2020), subsequently visualised in TreeViewer v. 2.0.1 (https:// treev iewer.org/) and edited in Affinity Publisher v. 2 (Serif (Europe), Nottingham, UK).The reference datasets, alignments and tree files were uploaded to the University of Pretoria research data repository hosted on Figshare (https:// www.doi.org/ 10. 25403/ UPres earch data.24004 071).

Results and discussion
Of the 426 genomes analysed in this study, 281 were correctly identified, 87 were misnamed, 12 were misidentified and 33 were submitted as Penicillium without a species name (see Table 1, Additional file 1: Table 1 and  Figs 1, 2).Of the correctly identified strains, 27 resolved in the P. camemberti species complex in the series Camembertiorum.This group is economically important and is typically used for the production of cheese like brie or camembert (Thom 1906).Taxonomically, this group and its six accepted species needs to be revised, but is complicated due to several past domestications (Ropars et al. 2020a, b).As there is little to no phylogenetic variation to guide identifications, we accept the name under which genomes from this group were submitted.Of the misidentified genomes, five belong to different genera including: GCA_023625675, which we believe to be a Candida species; GCA_023627405, which belongs to Aspergillus ustus; GCA_011750695, which belongs to Talaromyces minnesotensis; and GCA_002382835 and GCA_002382855, which belong to Talaromyces pinophilus.Six genomes were labelled with old names that have been synonymised, including: GCA_028828285 belonging to P. solitum (= P. majusculim) (Frisvad and Samson 2004)); GCA_025586815 belonging to P. desertorum (= P. glycyrrhizacola); GCA_015585885, GCA_015586035 and GCA_015585865 belonging to P. chrysogenum (= P. griseoroseum) (Houbraken et al. 2012); and GCA_028829675 belonging to P. glabrum (= P. tannophilum) (Houbraken et al. 2014).GCA_028974045 was submitted as a potential new species closely related to P. viridicatum and is identical to the recently described P. mali-pumilae (Hyde et al. 2019).Based on our analyses, we have identified three new species, including: GCA_028828675 in section Sclerotiora series Herqueorum; GCA_028827225 in section Fasciculata series Viridicata; and GCA_028826995, GCA_028974015 and GCA_028827235 in section Robsamsonia series Urticicola.Among the misidentified genomes were 12 that belong to different sections, including: GCA_000943775 and GCA_000943765 belonging to P. canescens in section Canescentia (not P. capsulatum in section Ramigena); GCA_015585765 and GCA_015585785 belonging to P. chrysogenum in section Chrysogena (not P. dipodomyicola in section Robsamsonia); GCA_028828875 and GCA_028826875   belonging to P. malacaense in section Idahoensia (not P. capsulatum in section Ramigena); GCA_015585975 belonging to P. rubens in section Chrysogena (not P. dipodomyicola in section Robsamsonia); GCA_020284065, GCA_019827435 and GCA_019828795 belonging to P. rubens in section Chrysogena (not P. fimorum in section Robsamsonia); GCA_015585905 belonging to P. rubens in section Chrysogena (not P. polonicum in section Fasciculata); and GCA_019804565 belonging to P. solitum in section Fasciculata (not P. robsamsonii in section Robsamsonia).We consider 87 genomes misnamed, with the submitted name being classified in the same series as our re-identified name.An example of this is the large number of genomes submitted as P. chrysogenum that belong to its closest relative, P. rubens in the series Chrysogena.
There are many reasons why genome sequences may have been submitted with names with which we disagree.The aim of this revision was not to criticise the submitters.Rather, we want to make our opinions known about the species to which available genomes belong, thus making the already very important resource that the submitters have created even more valuable.Based on our re-identifications, Penicillium genomes are currently available for 103 of 536 accepted species, representing both subgenera, 22 of 33 sections and 51 of 101 series.Introduction Fungi impact the breadth of biotechnology, animal and plant health, and food production.However, the knowledge of the contribution of molds in seasoned foods is still very limited compared to bacteria and yeasts (Tamang et al. 2016).Some species are used in the production of a particular fermented food, with significant contributions to improving food safety, nutritional value, organoleptic quality and contributing to food preservation (Bourdichon et al. 2012).Conversely, some molds are considered food contaminants causing spoilage and the production of toxic metabolites (EFSA 2023; Davies et al. 2021;Rico-Munoz et al. 2019;Avery et al. 2019).Mold growth is often considered an undesirable characteristic of aged products, with some exceptions, such as the use of Penicillium camemberti and Penicillium roqueforti in cheeses (Ropars et al. 2020a;Hymery et al. 2014), and Penicillium nalgiovense and Penicillium salamii on cured meats (Magistà et al. 2016;Mintzlaff and Leistner 1972).Some of the microbial species naturally found on fermented foods today originated from "domestication events".This process selected for the beneficial traits of wild species in some fermented food production, although this does not necessarily mean that they are suitable for other fermented foods (Bourdichon et al. et al. 2012;Kaur and Dua 2022).Besides P. nalgiovense and P. salamii, and according to the literature (Bourdichon et al. 2012;Alapont et al. 2014), we often isolated other Penicillium species during our investigations of the mold population on cured meat products, such as Penicillium biforme, P. brevicompactum, and P. solitum.
Penicillium biforme, has been hypothesized to have originated from the domestication of wild P. fuscoglaucum on cheese, in an episode independent of the one that led to the domestication of the sister species P. camemberti (Steenwyk 2019;Ropars et al. 2020b).Growth of P. biforme occurs more rapidly on salted than unsalted medium (Ropars et al. 2020b), and it is found on cheese rinds such as Tommes or fresh goat cheeses.It is also used to produce dry-cured meat (Ropars and Giraud 2022).Despite having a functional biosynthetic pathway, the lack of accumulation of ergot alkaloids, such as rugulovasines, in P. biformeaged cheese has been explained as nutrition-driven gene regulation preventing this fungus from producing ergot alkaloids in dairy products (Fabian et al. 2018).
Penicillium brevicompactum is often reported as a contaminant in spoiled dairy products (Garnier et al. 2017) and may contribute to the pleasant organoleptic characteristics of cured meats (Parussolo et al. 2019) due to its known lipolytic and proteolytic activity (Alapont et al. 2015).The main drawback to the presence of P. brevicompactum on cured meat, concerns the production of mycophenolic acid mycotoxin (Frisvad et al. 2004).
Penicillium solitum has been reported as a contaminant of cheese by several authors (Ramos-Pereira et al. 2019;Lund et al. 1995) but it is one of the predominant species in several dry-cured meat products, such as dry-cured sausages (Sørensen et al. 2008;Lopez Dıaz et al. 2001) and speck (Peintner et al. 2000).Although some studies include P. solitum among the producers of cyclopiazonic acid mycotoxin (Alapont 2014), it does not produce mycotoxins (Frisvad 2014) and its genome harbors only a partial cluster for the biosynthesis of patulin mycotoxin (Wu et al. 2019).It has been hypothesized that P. solitum could be responsible for the phenolic defect of hams, an unpleasant odor known to be caused by P. commune contamination (Scaramuzza et al. 2015), however it currently belongs to the list of seven Penicillium species reported with beneficial use for foods (Bourdichon et al. 2012).
Penicillium was amongst the predominant fungal genus capable of growth on the surface of dry fermented meats (Perrone et al. 2015;Magistà et al. 2017).Here we report the genome of three isolates of Penicillium from Italian salami, P. biforme ITEM 15300, P. brevicompactum ITEM 18316, and P. solitum ITEM 18327, respectively, which have been shown to be adapted to the curing conditions of fermented sausages production.

Materials and methods
The fungal isolates of Penicillium biforme ITEM 15300, P. brevicompactum ITEM 18316 and P. solitum ITEM 18327, were deposited as public resources at the Agri-Food Microbial Culture Collection-ITEM (http:// server.ispa.cnr.it/ ITEM/ Colle ction/).The genomic DNA was extracted from mycelium grown in Potato Dextrose Broth (Oxoid-Thermo Fisher, UK) in the dark at 25 °C, 200 rpm for 5 days.For RNA extraction each strain was inoculated as a spore suspension on Milk (50% v/v) agar, Brain Heart Infusion agar (Oxoid-Thermo Fisher, UK) and Yeast Extract Sucrose agar (Oxoid-Thermo Fisher, UK), and were grown in the dark at 25 °C for 5 days.The mycelia were recovered by vacuum filtration and ground in liquid nitrogen.Genomic DNA was isolated using the DNeasy Plant Mini Kit (Qiagen, Germany), and total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen, Germany), following the manufacturer's instructions.Quality and integrity of DNA and RNA was checked with NanoDrop 1000 (Thermo Fisher, UK) and Bioanalyzer 2100 (Agilent, Italy) analysis.
The draft genomes of ITEM 18316, ITEM 15300 and ITEM 18327 and the transcriptomic datasets for genome annotation, were generated with Illumina technology.For DNA library preparation, 200 ng of genomic DNA was sheared to 600 bp using a LE220 focused-ultrasonicator (Covaris, USA).The sheared DNA fragments were size selected by double-SPRI using TotalPure NGS beads (Omega Bio-tek, USA) and selected fragments were endrepaired, A-tailed, and ligated with Illumina compatible unique dual-index sequencing adaptors (IDT, USA).Libraries were sequenced on the NovaSeq 6000 (Illumina, USA) sequencing platform using NovaSeq XP v1 reagent kits (Illumina, USA), S4 flow cell, and following a 2 × 150 indexed run.
For the transcriptomic datasets, stranded cDNA libraries were generated using the Illumina Truseq Stranded mRNA Library Prep kit.Equimolar aliquots of RNA from each medium were pooled and used for library preparation.The mRNA fraction was purified from 1 ug of pooled total RNA using magnetic beads containing poly-T oligos, fragmented and reversed transcribed using random hexamers and SSII (Invitrogen, USA) followed by second strand synthesis.The prepared libraries were sequenced on the Illumina NovaSeq 6000 sequencing platform using NovaSeq XP v1 reagent kits (Illumina, USA), S4 flow cell, following a 2 × 150 indexed run.
Raw Illumina reads were filtered by Decontamination Using Kmers (BBDuk) in BBtools v. 38.79 (Bushnell 2020) which is capable of quality-trimming and filtering, adapter-trimming, and contaminant-filtering via kmer matching.Two million of the filtered genomic reads were subsampled to assemble the mitochondrial genome using GetOrganelle v. 1.7.1 (Jin et al. 2020).The nuclear genome assemblies were generated with SPAdes v. 3.14.1 (Bankevich et al. 2012), using a 20.0 M read-pair subsample of the resulting non-organellar reads, obtained by removing any organelle matching reads with BBtools v. 38.79 (Bushnell 2020).The filtered transcriptomic reads were used as input for de novo assembly of RNA contigs using Trinity v. 2.11.0 (Grabherr et al. 2011).
The genome assemblies were masked for repeats using RepeatMasker (Smit et al.1996(Smit et al. -2010) ) with the RepBase library 25.03 (Bao et al. 2015) and the most frequent repeats were identified by RepeatScout (Price et al. 2005).The completeness of the genome assemblies was performed with Benchmarking Universal Single-Copy Orthologs (BUSCO) v. 5.4.6, implemented in the Galaxy platform, using the eurotiales_odb10 lineage dataset (Manni et al. 2021).The nuclear genome was annotated using the JGI Annotation pipeline (Grigoriev et al. 2014) using a combination of ab initio, homology-based, and transcriptome-based gene models predicted from assembled RNA-sequencing data derived from the cultured fungus.
Predicted proteins were functionally annotated using SignalP v. 3 (Nielsen et al. 1997), TMHMM v. 2.0 (Melén et al. 2003), InterProScan v. 5.9-50.0(Quevillon et al. 2005), and BLASTp alignments against the NCBI NR, SwissProt, KEGG (Kanehisa et al. 2006), and KOG (Koonin et al. 2004) databases.Transcription factors were assigned based on Pfam domains.Gene ontology (GO) terms (Ashburner et al. 2000) were assigned based on InterPro and SwissProt hits.Protein alignments by BLASTp against TCDB were used for transporter classifications (Saier et al. 2016) and MEROPS for peptidase classifications (Rawlings et al. 2014).CAZymes were annotated as described in Lombard et al. (2014).Secondary metabolite clusters and classifications were inferred from Pfam domain content and physical proximity based on the SMURF algorithm (Khaldi et al. 2010).Cytochrome 450 (CYP) subfamilies were assigned based on HMMs, which were labelled based on homology to manually curated CYP genes (Nelson 2009).The HMMs corresponding to different CYP subfamilies were derived from CYP sequences in NCBI NR and Myco-Cosm (~ 220 K sequences) followed by UCLUST clustering (Edgar 2010) and iterative HMM building.Finally, biosynthetic gene clusters (BGCs) were identified using anti-SMASH fungal v. 6.1.1 (Blin et al. 2021) with default parameters.
The taxonomic identity of the species was confirmed by phylogenetic analysis of four combined gene regions.Partial DNA sequences from the internal transcribed spacer (ITS) region, calmodulin (CaM) gene, beta-tubulin (BenA) gene, and DNA-dependent RNA polymerase II second largest subunit (RPB2) gene, were extracted from the genomes.The nucleotide sequences of Penicillium type strains in sections Brevicompacta and Fasciculata were retrieved from GenBank, following Houbraken et al. (2020).The sequences were aligned using the online version of MAFFT v. 7. (Katoh et al. 2019).IQ-TREE v. 2.2.0 (Minh et al. 2020) implemented ModelFinder (Kalyaanamoorthy et al. 2017) to calculate the best-fit model according to the Bayesian Information Criterion (BIC) score on the partitioned dataset (Chernomor et al. 2016), and infer the Maximum Likelihood phylogenetic tree based on 10,000 ultrafast bootstrap support (Hoang et al. 2018).taxonomy of the camemberti clade has not been fully resolved with potent genetic markers, thus several misidentified isolates have recently been reassigned to the right species using whole-genome-based analyses (Ropars et al. 2020b).The P. biforme ITEM 15300 clusters with the type species of P. biforme with high bootstrap support.The availability of these three genomes of Penicillium species from dry-cured meat environments will allow for new comparative studies for species adapted to this challenging but economically important environment. Authors

Draft genome sequence of Penicillium cvjetkovicii isolated from Italian salami Introduction
The ascomycete genus Penicillium is home to tremendous biochemical and enzymatic diversity that contributes to its phenotype and may also have use in biotechnology.
Penicillium cvjetkovicii belongs to the slow-growing subgenus Aspergilloides, section Cinnamopurpurea, series Idahoensia (Houbraken et al. 2020).Species from this section have been found as contaminants on food products, such as P. fluviserpens isolated from tomato fruit (Nguyen et al. 2020), P. cinnamopurpureum and P. malacaense reported from meju samples in Korea (Jung et al. 2012;Kim et al. 2015), and P. cinnamopurpureum contributing to the potato-taste-defect of coffee (Hale et al.

2022).
Penicillium cvjetkovicii has been described relatively recently (Peterson et al. 2015).A cheese isolate was originally obtained from Spain (Marin et al. 2014) and identified by Peterson et al. (2015) as P. cvjetkovicii, although it had initially been considered to be P. chermesinum (Marin et al. 2014, Peterson et al. 2015).Morphologically, this fungus is characterized by the monoverticillata penicilli and the production of vinaceous to reddish-brown soluble pigments.The production of the vinaceous to reddish-brown soluble pigments, which is typically observed in most of the species belonging to section Cinnamopurpurea (Peterson et al. 2015), was probably related to the observation of small dark spots on the surface of aged cheeses (Ramos-Pereira et al. 2019).Two species in this section, P. colei and P. monsserratidens produce citreoviridin, a mycotoxin synthesized by highly reducing polyketide synthases.Similar to other Penicillium species in this section, however, P. cvjetkovicii is not known to produce mycotoxins or other toxic metabolites (Peterson et al. 2015).Other species in section Cinnamopurpurea are hypothesized to be beneficial in developing the unique characteristics of typical foods, such as, P. gravinicasei recently isolated from cheese (Anelli et al. 2018), and P. nodulum which participates in the softening of cassava tissue during the fermentation of cassava dough into agbelima (Amoa-Awua 1997).
Very few species belonging to other sections are associated with cheese.This is the case of P. glabrum (section Aspergilloides), the recently described P. cvjetkovicii (section Cinnamopurpurea) and P. citrinum (section Citrina) (Houbraken et al. 2020).
Here we report the genome of an isolate of P. cvjetkovicii, which was isolated from a dry cured sausage and tested as a surface mold starter to produce the product.

Nucleotide sequence accession number
The genome assembly and annotations are available from the JGI Fungal Genome Portal MycoCosm (Grigoriev et al. 2014; https:// mycoc osm.jgi.doe.under JGI Project ID 1289847 and has been deposited at GenBank under BioProject number PRJNA971650(BioSample n.SAMN35051277; Project Accession n.SRP442271).
For genomic DNA extraction, the mycelium was grown in potato dextrose broth (PDB) in the dark at 25 °C, 200 rpm for 5 d.The mycelium was recovered by vacuum filtration and ground in liquid nitrogen.The DNeasy Plant Mini Kit (Qiagen, Germany), was used to extract genomic DNA following the manufactures' instructions.For RNA extraction, the strain was inoculated as a spore suspension on milk (50% v/v) agar, brain heart infusion agar (Oxoid, UK) and yeast extract sucrose (YES) agar and grown in the dark at 25 °C for 5 days.The mycelia was collected, ground in liquid nitrogen and total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen, Germany), according to the manufactures' instructions.Quality and integrity of DNA and RNA was checked with NanoDrop and Bioanalyzer 2100 analysis.
The draft genome of P. cvjetkovicii ITEM 18317 and the transcriptomic datasets for genome annotation were generated with Illumina technology.For DNA library preparation, genomic DNA was sheared using a Covaris LE220 focused-ultrasonicator.The sheared DNA fragments were used for library preparation and sequenced on the Illumina NovaSeq 6000 sequencing platform 2 × 150 indexed run.Raw Illumina reads were filtered for quality and artifacts using BBTools software package v. 38.79 (Bushnell 2020).Mitochondrial genome was assembled using get_organelle v1.7.1 (Jin et al. 2018).The genome assembly was generated with SPAdes v3.14.1 (Bankevich et al. 2012) filtering out the resulting organellar reads.For the transcriptomic dataset, stranded cDNA libraries were generated and sequenced on the Illumina NovaSeq 6000 sequencing platform 2 × 150 indexed run.The raw Illumina reads were filtered and trimmed for quality and artifacts using BBTools software package v. 38.79 (Bushnell 2020) and used as input for de novo assembly of RNA contigs using Trinity (v2.11.0) (Grabherr et al. 2011).
The genome assembly was masked for repeats using RepeatMasker (Smit et al.1996(Smit et al. -2010) ) with the RepBase library v25.03 (Jurka et al. 2005) and the most frequent repeats were identified by RepeatScout (Price et al. 2005).The genome assembly completeness was estimated with BUSCO v5.4.6 (lineage dataset: eurotiales_odb10), implemented in the Galaxy platform (Manni et al. 2021).The nuclear genome was annotated with the JGI Annotation pipeline (Grigoriev et al. 2014), using a combination of ab initio, homology-based, and transcriptome-based gene predictors.Predicted proteins were functionally annotated using SignalP v3 (Nielsen et al.1997), TMHMM v2.0 (Melén et al. 2003), InterProScan 5.9-50.0(Quevillon et al. 2005), and BLASTp alignments against the NCBI NR, SwissProt, KEGG (Kanehisa et al. 2006), and KOG (Koonin et al. 2004) databases.Transcription factors were assigned based on Pfam domains.Gene ontology (GO) terms (Ashburner et al. 2000) were assigned based on InterPro and SwissProt hits.Protein alignments by BLASTp against TCDB were used for transporter classifications (Saier et al. 2016) and MEROPS for peptidase classifications (Rawlings et al. 2014).CAZymes were annotated as described in Lombard et al. (2014).Secondary metabolite clusters and classifications were inferred from Pfam domain content and physical proximity based on the SMURF algorithm (Khaldi et al. 2010).Cytochrome 450 (CYP) subfamilies were assigned based on HMMs, which were labelled based on homology to manually curated CYP genes (Nelson 2009).The HMMs corresponding to different CYP subfamilies were derived from CYP sequences in NCBI NR and Myco-Cosm (~ 220 K sequences) followed by UCLUST clustering (Edgar 2010) and iterative HMM building.Finally, biosynthetic gene clusters (BGCs) were identified using anti-SMASH fungal v. 6.1.1 (Blin et al. 2021) with default parameters.
The taxonomic identity of P. cvjetkovicii ITEM 18317 was confirmed with phylogenetic analysis of three combined gene regions.Partial DNA sequences from the internal transcribed spacer (ITS) region, calmodulin (CaM) gene, and beta-tubulin (BenA) gene, were extracted from the genome.The nucleotide sequences of Penicillium ex-type strains in the same section Cinnamopurpurea were retrieved from GenBank, following Houbraken et al. (2020).The sequences were aligned using the online version of MAFFT v. 7. (Katoh et al. 2019).IQ-TREE 2.2.0 (Minh et al. 2020), implemented with ModelFinder (Kalyaanamoorthy et al. 2017), was used to calculate the best-fit model according to the Bayesian Information Criterion (BIC) score on the partitioned dataset (Chernomor et al. 2016), and infer the Maximum Likelihood phylogenetic tree based on 10,000 ultrafast bootstrap support (Hoang et al. 2018).

Results and discussion
Fermented meat products represent a significant part of the Mediterranean diet and their production and commercialization contribute to the local economies (Baka et al. 2011).In the last years, many studies have focused on the characterization of fermented sausages microbiota (Ferrocino et al. 2018;Magistà et al. 2016;Perrone et al. 2015).Many fungal species are well adapted to the ecological conditions encountered during the meat fermentation.That is why many fermented meat products, including fermented sausages, are characterized by yeasts and molds growth on the casing surface.The mycobiota of traditional dry-cured meat products is usually characterized by the presence of mould species belonging to the Penicillium genus, mainly P. solitum, P. nalgiovense, P. chrysogenum, P. olsonii, P. commune, and P. salamii (Magistà et al. 2017).Besides P. nalgiovense, other mould species might be isolated from fermented meat products, probably transferred to the casing surface through airborne contamination or through the use of spices or salt.This is the case of P. cvjetkovicii ITEM 18317 isolated from an Italian salami.The draft genome of strain ITEM 18317 was generated at the DOE Joint Genome Institute (JGI) using Illumina short read sequencing technology.A total of 52,369,604 raw reads were generated, yielding 51,612,264 reads (7.74 Gb) after filtering.For transcriptome sequencing, 180,428,788 reads were generated, of which about 30% were discarded after filtering.Single copy BUSCOs 4125

Missing BUSCOs 56
Total BUSCO groups searched The draft genome assembly process yielded 272 contigs and 202 scaffolds with a genome size of 25.62 Mbp (Table 5).Additionally, one contig was generated for the mitochondrial genome with a length of 27.19 Kbp.Phylogenetic analysis confirmed the identity of the strain P. cvjetkovicii ITEM 18317 (Fig. 4).
Gene prediction produced 9657 protein-coding gene models, of which 7529 were annotated with InterPro (Table 6).Annotation of secondary metabolites performed with anti-SMASH confirmed the absence of mycotoxin-producing biosynthetic gene clusters in P. solitum ITEM 18327.Here we present the first draft genome of P. cvjetkovicii.Its availability for the scientific community will facilitate the investigation of the biology of this fungus, given our hypothesized application as a fungal starter for meat fermentation.

Draft genome assemblies of four Pewenomyces species from Chile Introduction
Pewenomyces is a recently described genus in Coryneliaceae (Eurotiomycetes, Coryneliales) that contains four species: Pew.kutranfy, Pew.lalenivora, Pew.tapulicola, and Pew.kalosus (Balocchi et al. 2021(Balocchi et al. , 2022)).All these fungi were discovered on Araucaria araucana trees in Chile, associated with cankers on branches and young stems.Pewenomyces kutranfy, the type species for the genus, was confirmed to be pathogenic and the causal agent of the cankers observed in these trees (Balocchi et al. 2021).The lifestyle of the other three species remains uncertain, but they are most likely saprophytes and/or endophytes (Balocchi et al. 2022).The origin of these Pewenomyces species is unknown, although it has been suggested that they are native to the Chilean mountain ranges where they were discovered (Balocchi et al. 2022).This suggeston was based on their temperature preference for growth, the unique and harsh niche where they are found, the species diversity associated with a single host species, and the diversity of Coryneliaceae found in Chilean native forests (Fitzpatrick 1942;Butin 1970;Benny et al. 1985aBenny et al. , 1985b;;Balocchi et al. 2022).Determining the origin and diversity of these fungi, which is particularly relevant for the emergent pathogen Pew.kutranfy, would require research-based evidence.Studies of this nature have not been performed for any other members of the Coryneliaceae.
The Coryneliaceae is a relatively small fungal family containing approximately 60 species distributed in nine accepted genera.The majority of the species were described before DNA-based techniques were routinely used for taxonomy (Fitzpatrick 1920), and the identity of more than half of the accepted species (including three whole genera) still need to be validated using phylogenetic analyses.Some of the most well-known species in the family are tree pathogens, including Caliciopsis pinea on Pinus spp. in the USA (Munck et al. 2015), Hypsotheca pleomorpha on Eucalyptus spp. in Australia (Pascoe et al. 2018), and Corynelia spp. on Podocarpaceae in South Africa (Wood et al. 2016).Similar to Pewenomyces spp., these fungi are mostly assumed to be native where they have been found, a premise based on their occurrence on tree species native to the area and/or apparently narrow geographical distributions (Wood et al. 2016;Pascoe et al. 2018;Migliorini et al. 2020).However, no studies have considered the diversity or biology of these fungi using DNA-based techniques.The genomes of two species of the family, C. pinea and C. orientalis, are available in open collections (e.g.JGI, GenBank), but no published studies have emerged from these resources.Sequencing the genomes of species in Pewenomyces creates further opportunities for studying the biology and evolution of fungi in this relatively small and unique group of fungi, which occur in unique niches distributed around the world.

Materials and methods
Cultures of all isolates were obtained from the culture collection (CMW) of the Forestry and Agricultural Biotechnology Institute (FABI), University of Pretoria, South Africa.Isolates were grown in 25 mL glass vials containing ~10 mL of liquid YM broth media (2% malt extract and 0.2% yeast extract) for 4-7 d in the dark with continuous shaking.Mycelia were collected into 2 mL Eppendorf tubes by straining it through sterile gauze, freeze dried, and then ground with metal beads using a mixer mill (MM 301, Retsch GmbH; 30 oscillations/s for 3 min).DNA was extracted using a salt-extraction protocol described by Aljanabi and Martinez (1997) and modified by Duong et al. (2013).Resulting DNA quality and quantity was assessed by gel electrophoresis (1% agarose gel; 12 min at 110 V) and with a Qubit 4 fluorometer (Qubit Assays, Termo Fischer).Illumina whole genome sequencing was carried out by Macrogen where a library with a 350 bp insert size was prepared and sequenced on the NovaSeq 6000 platform to generate 151 bp pairedend reads.
The resulting Illumina reads were trimmed and assembled using CLC Genomics Workbench v 22.0.4(QIA-GEN, Aarhus) with default parameters, discarding contigs shorter than 500 bp.The completeness of the genome was assessed using Benchmarking Universal Single-Copy Orthologs (BUSCO) v. 5.3.2 with the euro-tiomycetes_odb10 dataset (Manni et al. 2021).Genome metrics including N50, L50, GC content, and the genome size were obtained using QUAST v 5.0.2 (Mikheenko et al. 2018).The number of protein-coding genes in each genome was predicted using AUGUSTUS v. 3. 4.0 (Keller et al. 2011) based on the gene models of Aspergillus fumigatus.
The identity of the isolates sequenced in this study were verified by conducting phylogenetic analyses with sequences of the ITS, nc LSU rDNA and RPB2 gene regions extracted from the resulting assemblies.These analyses were performed by including these sequences in modified datasets obtained from Balocchi et al. (2022), which included representative sequences of most genera in the Coryneliaceae.The datasets for each gene region were compiled and aligned using MAFFT (Katoh et al. 2017).Assembled datasets were edited and concatenated using MEGA XI (Tamura et al. 2021) and maximumlikelihood trees were built for the concatenated datasets using the IQ-TREE Web server (Trifinopoulos et al. 2016).Evolutionary models for the analyses were selected using ModelFinder (Kalyaanamoorthy et al. 2017) and statistical support was calculated with Ultrafast Bootstrap analysis (Minh et al. 2013).Resulting trees were visualized and edited using FigTree v.1.4.4 (http:// tree.bio.ed.ac.uk/ softw are/ figtr ee/) and edited using Affinity Designer v.1.10.5.1342 (Serif, Nottingham, UK).

Results and discussion
Illumina sequencing resulted in 33-39 million reads per isolate, of which between 99.4% and 99.7% remained after trimming.After assembly and filtering, the number of scaffolds per genome varied between isolates of the different species (Table 7).The lowest number of scaffolds was obtained for Pew.kutranfy CMW54240 [n = 203; N 50 = 510 kb], and the largest number of scaffolds were for Pew.lalenivora isolates CMW54250  (Balocchi et al. 2021(Balocchi et al. , 2022)).Phylogenetic analyses with the individual and concatenated datasets consistently resolved the genome sequenced taxa in the Pewenomyces clade next to their corresponding Sanger-based equivalent (Fig. 5), confirming the identity of the sequenced isolates.The genomes produced in this study will provide a valuable future resource for comparative studies involving species in Pewenomyces and Coryneliaceae.This includes studies ranging from phylogenomics to genome comparisons for biological and ecological questions, mating type system identification, and the development of molecular tools such as microsatellite markers for population genetic studies.
Authors: Felipe Balocchi*, Irene Barnes, Brenda D. Wingfield, Anja Piso, Tuan A. Duong* *Contact: Tuan.Duong@fabi.up.ac.za; felipe.balocchi@fabi.up.ac.za.Andjic et al. 2019).Teratosphaeria species can be found in asymptomatic Eucalyptus trees (Kemler et al. 2013;Marsberg et al. 2014), and some only cause mild disease symptoms (Hunter et al. 2011).In contrast, a group of closely related Teratosphaeria species with Kirramyces asexual morphs are aggressive pathogens and result in severe TLB disease on Eucalyptus trees established in plantations, predominantly in areas having tropical and subtropical climates (Andjic et al. 2019).These include species such as T. destructans, T. eucalypti and T. pseudoeucalypti.Most recently, T. carnegiei has been described residing in this group of cryptic species (Crous et al. 2022).

IMA GENOME-F 18E
Teratosphaeria carnegiei was discovered amongst a collection of isolates thought to be those of T. pseudoeucalypti isolated from TLB symptoms in a Eucalyptus grandis x E. camaldulensis plantation in New South Wales (NSW), Australia (Aylward et al. 2021).The isolates were assessed using a microsatellite panel designed to identify TLB species (Havenga et al. 2020) and most isolates were identified as T. pseudoeucalypti.However, two of these isolates had genotypes distinct from those of any other TLB pathogens.Phylogenetic analyses showed that these two isolates resided in a monophyletic group with other isolates that had previously been recognized as variants of T. eucalypti (Andjic et al. 2010;Crous et al. 2022), but were distinct from both T. eucalypti and T. pseudoeucalypti.
Teratosphaeria carnegiei appears to be of minor economic significance as a pathogen.It has been discovered only twice, both times in northern NSW as part of population-level isolations of T. eucalypti or T. pseudoeucalypti (Andjic et al. 2010;Crous et al. 2022).It's low frequency of isolation and co-occurrence with aggressive pathogens raises the question as to whether it can cause disease independently.However, its position as the species most closely related to two damaging TLB pathogens, makes it of considerable interest.This prompted the present study to sequence the genome T. carnegiei in order to compare it with other species causing severe TLB.

Nucleotide accession number
The genomic sequences of T. carnegiei have been deposited at DDJ/EMBL/GenBank under the accession JANYMD000000000.This paper describes the first version.

Material and methods
The culture of T. carnegiei CMW 52470 was obtained from the culture collection of the Forestry and Agricultural Biotechnology Institute (FABI) at the University of Pretoria and grown on malt extract agar (Merck, Wadeville, South Africa) at room temperature for approximately two weeks.DNA extraction proceeded as previously described for Teratosphaeria species (Wingfield et al. 2019).Sequencing took place at the Central Analytical Facilities (CAF), Stellenbosch University, using the Ion S5 ™ System and an Ion 530 ™ Chip (Thermo Fisher Scientific, MA, USA), at a target read length of 600 bp.After assessing read quality with FastQC 0.11.9 (Andrews 2010), the genome was assembled with SPAdes 3.15.2(Bankevich et al. 2012), using the built-in read trimming function and kmer values of 21, 33, 55 77, 99 and 127.Genome completeness was assessed with BUSCO 4.1.4 (Simão et al. 2015), genome coverage was estimated by aligning the reads back to the genome with Bowtie 2.4.1 (Langmead and Salzberg 2012) and contamination was assessed with BlobToolKit 1.2 (Challis et al. 2020).Repeat content was determined with RepeatModeler 2.0.3 (Flynn et al. 2020) and open reading frames were predicted with the Funannotate 1.8.12 predict pipeline (Palmer and Stajich 2020).
The phylogenetic position of the sequenced strain relative to the other known T. carnegiei isolates and closely related Teratosphaeria species was determined using the ITS and beta-tubulin regions.The Maximum Likelihood tree was constructed by aligning the sequences with MAFFT v7.490 (Katoh and Standley 2013), manual alignment trimming and using ModelTest-NG 0.1.6(Darriba et al. 2020) to identify the best nucleotide substitution model.Individual and concatenated gene trees were determined with RAxML-NG 1.1 (Kozlov et al. 2019), applying the transfer (TBE) bootstrap support of (Lemoine et al. 2018).

Results and discussion
Sequencing yielded 11.7 million reads ranging between 25 and 840 bp (mode = 532 bp) and FastQC did not flag any low-quality or overrepresented sequences.The final 27.69 Mb assembly had a coverage of approximately 150 X and comprised 1,135 contigs > 1 kb, with an L50 of 71 and an N50 of 128,647 bp.Genome completeness according to the Fungi_odb10 dataset was estimated at above 98% (745 complete BUSCOs = 98.3%) and BlobToolKit did not detect significant contamination.Funannotated predicted 9,464 protein-coding and 57 tRNA genes.
The 7.29% repetitive sequences identified in the T. carnegiei genome likely contributed to the low assembly contiguity.This proportion was less than half of the ca.16-17% estimated for the assemblies of T. destructans CMW 44962 (Wingfield et al. 2018) and T. eucalypti CMW 54005 (Aylward et al. 2022), the two other Teratosphaeria species sequenced with the same technology.The T. carnegiei assembly, however, had better N50 and L50 values than either of those assemblies, further implying that the repeat content influenced the continuity of the assembly.The lower repeat content also influenced assembly size as the T. carnegiei genome was more than 2 Mb smaller than those of T. destructans CMW 44962 and T. eucalypti CMW 54005.
Phylogenetic analyses of the ITS and beta-tubulin regions placed T. carnegiei within the lineage of tropical and subtropical leaf pathogens, where it shares a wellsupported (91%) clade with T. eucalypti and T. pseudoeucalypti (Fig. 6).The relationship among these three cryptic species remains to be resolved, although the analysis of Andjic et al. (2010) suggests that T. carnegiei and T. eucalypti are sister species.All three taxa are known from diseased trees in eastern Australia plantations, but T. eucalypti and T. pseudoeucalypti are also known to cause disease problems beyond this range.For example, T. eucalypti is well -known in New Zealand (Hood et al. 2002) and T. pseudoeucalypti is important pathogen in South America (Cândido et al. 2014;Soria et al. 2014;Ramos and Pérez 2015).In contrast, the four T. carnegiei strains included in Fig. 6 are the only known isolates of this species, representing samples taken in 2009 (MUCC strains) and 2018 (CMW strains) from two plantations approximately 50 km apart (Andjic et al. 2010).A single point mutation in the beta-tubulin gene separates the isolates from these two sites.
The genome sequence of T. carnegiei brings the total number of sequenced Teratosphaeria species to nine.In addition to the aggressive tropical and subtropical foliar pathogens and the stem canker pathogens included in Figure 6, the species for which genomes have been sequenced include T. nubilosa which is an important pathogen of cold-tolerant Eucalyptus species such as Introduction Trichoderma is a genus of mainly asexual fungi belonging to the Hypocreaceae, primarily isolated from soils, roots, or leaves of plants present in every type of soil (tropical and temperate) (Howell 2002).These filamentous fungi present high genetic diversity and can be used to produce various products of commercial and ecological interest (Gupta et al. 2014).The benefits of Trichoderma species are well described in many sectors of industry and agriculture (Gupta et al. 2014).
Trichoderma species exert biocontrol against fungal phytopathogens using several mechanisms.Trichoderma can attack phytopathogens directly, using mechanisms such as mycoparasitism and antibiosis, or indirectly, competing for nutrients and space, or promoting plant growth and defense mechanisms (Sood et al. 2020;Vinale et al. 2008).Trichoderma's most salient characteristic is their ability to parasitize other fungi, which is ensured by a broad range of molecules, especially cell wall degrading enzymes (CWDEs) (Sood et al. 2020).Initially, Trichoderma uses transporters like the tripeptide transporter and the ABC transporter, to move towards a phytopathogenic fungus (Chet et al. 1981).Subsequently, Trichoderma produces innumerous CWDEs that hydrolyze the cellular walls of phytopathogenic fungi, ultimately leading to their death.Among CWDEs are chitinases, endochitinases, xylanases, proteases, and β-glucanases (Sharma et al. 2011).Trichoderma species, also present an antifungal arsenal that includes terpenes, pyrones, gliotoxin, gliovirin, and peptaibols, with activity against phytopathogens (Sharma et al. 2019;Vinale et al. 2008Vinale et al. , 2020)).When grouped together, antifungal molecules and CWDEs enhance their antibiotic effect against a broad spectrum of fungal phytopathogens (Tronsmo 1991).Trichoderma species can stimulate plant defenses, using molecules recognized as elicitors by the plant to trigger systemic defences (Hermosa et al. 2012;Lazazzara et al. 2021).Organic volatile compounds (VOCs), secondary metabolites in low concentrations, and phytohormonelike compounds produced by Trichoderma species can induce plant defenses, mainly salicylic acid, and ethylene-dependent defences (Hermosa et al. 2012;Lazazzara et al. 2021).Trichoderma can also increase plant root growth and productivity by influencing plant hormonal balance, increasing plant nutrient uptake, and solubilizing soil nutrients (Pozo et al. 2002;Sood et al. 2020).However, it is still unknown how these processes occur at a molecular level.
Trichoderma atroviride SC1 biocontrol potential against grapevine pathogens, responsible for several important diseases (i.e.grapevine trunk diseases or downy mildew), is well documented (Berbegal et al. 2020;Lazazzara et al. 2021;Leal et al. 2021;Martínez-Diz et al. 2021;Pertot et al. 2017).In this study, we present the draft genome sequence of Trichoderma atroviride SC1 with the aim of advancing knowledge about this strain and its biocontrol potential against grapevine diseases.

Nucleotide sequence accession number
The draft genome of Trichoderma atroviride SC1 CBS strain 122089 reported here is made of high-quality assemblies.It has been deposited in GenBank under Acc.No. JAQOTD000000000 (BioProject No. PRJNA923860, assembly No. GCA_028554805.1, biosample No. SAMN32746547).

Materials and methods
The strain was cultivated from the commercial product Vintec ® (Belchim crop Protection, Londerzeel, Belgium), was purified by single-spore isolation and maintained on potato dextrose agar (PDA) medium at 25 °C in the darkness.DNA was extracted with NucleoSpin Tissue (Macherey-Nagel, Duren, Germany), following the manufacturer's protocol.Firstly, the complete ITS region, including the 5.8S gene, were amplified with ITS1/ ITS4 (White et al. 1990), using the amplicon sequencing according to Eichmeier et al. (2010).The same DNA was used for genome library construction with the Nextera XT DNA Library Preparation Kit (Illumina, San Diego, USA).The library was sequenced using Min-iSeq High Output Reagent Kit (300-cycles) (Illumina) with 2 × 150PE read option.The same DNA sample was sequenced using the Oxford Nanopore (LP-150), Grid-ION FC (Oxford Nanopore Technologies, Oxford, UK), single-end, 1-200 kb reads, 5-10 Gb (DS-210).The sequence quality was checked using the FastQC-0.10.1 program (Andrews 2010).A FASTX-Toolkit Clipper (http:// hanno nlab.cshl.edu/ fastx_ toolk it/), specifying the Q33 parameter, was used to remove adaptors, and low-quality reads were discarded.Contigs of individual reads were assembled de novo using SPAdes genome assembler v. 3.15.2 (Prijibelski et al. 2020) with default settings, and a hybrid assembly of Illumina and nanopore reads was performed.The ab initio gene prediction was performed using Augustus (Keller et al. 2011) (-species = botrytis_cinerea -strand = both) for the assembled genome of T. atroviride SC1, resulting in predicted coding sequences.BUSCO 5.2.2 (Manni et al. 2021) revealed complete and single-copy proteins, posteriorly identified according to their function.Carbohydrate-active enzymes (CAZymes) were predicted using CAT and dbCAN3 servers (Yin et al. 2012).Signal peptides were detected by HMMER (Zhang and Wood 2003).Annotation was performed using JGI (Join Genome Institute).The search for secondary metabolite clusters was done using JGI MycoCosm.Placement of T. atroviride SC1 within the closest Trichoderma species (Trichoderma Viride clade) was verified using phylogenetic analysis of a ITS region.The dataset was aligned using the MAFFT v. 7 using the European Bioinformatics Institute platform (EMBL-EBI, https:// www.ebi.ac.uk).Obtained alignment was manually checked and edited using Geneious Prime ® 2023.1.1 (Biomatters, Inc., New Zealand).The maximum likelihood (ML) tree was constructed using IQ-TREE 2 (Minh et al. 2020).The best models for ML analyses were selected based on the Akaike Information Criterion (AIC) calculated in IQ-TREE 2. Trees were visualized in FigTree v. 1.4.4 and edited in Adobe Illustrator CC 2019.

Results and discussion
Using Oxford Nanopore technology 1,503,165 reads were obtained with mean read length 4,918 bp.Sequencing by synthesis provided 14,630,016 reads and 13,771,719 reads passed the chastity filter.Genome coverage reached 50.5×.De novo assembly of T. atroviride SC1 CBS 122089 resulted in a genome size 35,757,960 bp with G + C content of 49.86%, and 603 contigs, with a scaffold length in which 50% of the total assembly length are covered (N50) values of 312,579 bp and the number contigs whose summed length is N50 (L50) of 35.The sequencing of ITS region (submitted to GenBank Acc.No. OP618118) confirmed a similarity score of 100% with T. atroviride available accessions, 545/545 nts.The phylogenetic placement of the genome is provided in Fig. 7. Genome completeness was estimated to be 97.2%corresponding to 96.8% complete and single-copy BUSCOs, 0.4% complete and duplicated BUSCOs and 2.2% missing BUSCOs.A total of 11,401 gene models were predicted in the T. atroviride SC1 assembly.Eighty-six signal peptides were detected by HMMER using dbCAN3.Signal peptides act as a zip codes, marking the protein secretion pathway as well as protein target location.In addition to protein targeting, a number of critical functions with or without regard to the passenger proteins have been attributed to signal peptides (Owji et al. 2018).A total of 129 CAZyme subfamilies were detected in 443 contigs using HMMER.The most represented CAZymes belonged to the subfamily (SBFs) GH18.Further classification of CAZymes based on their catalytic activity showed a high proportion of glycoside hydrolases (62 SBFs-48.1%),glycosyl transferases (30 SBFs-23.3%),carbohydrate-binding molecules (13 SBFs-10.1%),auxiliary activities (11 SBFs-8.5%),carbohydrate esterases (8 SBFs-6.2%),polysaccharide lyases (5 SBFs-3.9%).Compared to T. afroharzianum T11-W, T. harzianum CBS 266.95, T. pleuroticola (Zhou et al. 2020), or even T. atroviride IMI 206040 (Kubicek et al. 2011), T. atroviride SC1 has a high proportion of glycoside hydrolases.Using Micro-Station Reader BioTek ELx808BLG (Biolog) and carbon sources (CS) in FF MicroPlate (Biolog Inc.), consumption was detected of 64 CS by T. atroviride SC1.This fungus was clearly identified as T. atroviride according to the FF MicroPlate database of Biolog Inc. Secondary metabolites are essential for fungal growth and development, providing protection against various stresses (Calvo et al. 2002).The search for secondary metabolite clusters revealed the presence of 38 clusters 12 × type I polyketide synthase, 11 × non-ribosomal peptide synthetase fragment, 8 × non-ribosomal peptide synthetase, 3 × terpene, 2 × polyketide-like and 2 × hybrid clusters.
In addition to T. atroviride, the genomic resource presented here includes seven other genomes (available from the National Center for Biotechnology Information) for this species associated with an effective biocontrol properties.The comparison of the available T. atroviride genome assemblies (Table 8) shows that the strain SC1 has the smallest genome and comparing to IMI 206040 and P1 strains has lower number of gene models.The availability of genomic resources for these

Fig. 1
Fig. 1 Phylogenetic tree of Penicillium subgenus Aspergilloides based on a concatenated dataset of BenA, CaM, Cct8, RPB1, RPB2 and Tsr1.Data obtained from available NCBI genomes appear in bold black text.Reference sequences appear in grey text.Ex-type strains are indicated by superscript T. The tree was rooted to P. expansum.

Fig. 2
Fig. 2 Phylogenetic tree of Penicillium subgenus Penicillium based on a concatenated dataset of BenA, CaM, Cct8, RPB1, RPB2 and Tsr1.Data obtained from available NCBI genomes appear in bold black text.Reference sequences appear in grey text.Ex-type strains are indicated by superscript T. The tree was rooted to P. glabrum.

PFig. 3
Fig.3ML phylogenetic tree of Penicillium sections Fasciculata and Brevicompacta indicating the three sequenced isolates (in blue), and ultrafast bootstrap support at the nodes.

Fig. 5
Fig. 5 Maximum likelihood tree for the concatenated ITS, nc LSU rDNA and RPB2 for members of the Coryneliaceae.Sequences extracted from genomes produced in this study are highlighted in red.Numbers on branches indicate Bootstrap values (n = 1000).

Fig. 6
Fig. 6 Maximum likelihood phylogeny of the concatenated ITS and beta-tubulin regions showing the phylogenetic position of Teratosphaeria carnegiei relative to other leaf pathogens in the tropical/subtropical clade.The stem pathogens T. gauchensis and T. zuluensis have been used as outgroups.Values on branches represent the transfer (TBE) bootstrap support.Superscripts indicate ex-type (ET), reference (R) and genome (g) strains.The strain sequenced in this study is shown in bold.GenBank accession numbers are available in Quaedvlieg et al. (2014), Aylward et al. (2019) and Crous et al. (2022).

Fig. 7
Fig. 7 Maximum likelihood tree based on ITS region.Values at branch nodes are the bootstrapping confidence values with those ≥75% shown.The Trichoderma atroviride SC1 isolate sequenced in this study is indicted in red

Table 1
Summary of genomes re-identified during this study.See Additonal file 1: Table1for the full list of strains analysed during this study.

Table 2
Whole genome assembly features of the three Penicillium species isolated from Italian salami.

Table 3
Completeness of the genome assemblies of the three Penicillium species isolated from Italian salami calculated with BUSCO.

Table 4
Details of coding genes, exons, and introns of the three Penicillium species isolated from Italian salami.

Table 5
Whole genome assembly features of Penicillium cvjetkovicii ITEM 18317.

Table 6
Details of coding genes, exons and introns in P. cvjetkovicii ITEM 18317 genome.

Table 7
Genome assembly statistics for six isolates representing four Pewenomyces species.

Table 8 -
see additional TABLE Authors: