IMA Genome-F 6: Draft genome sequences of Armillaria fuscipes, Ceratocystiopsis minuta, Ceratocystis adiposa, Endoconidiophora laricicola, E. polonica and Penicillium freii DAOMC 242723.

The genomes of Armillaria fuscipes, Ceratocystiopsis minuta, Ceratocystis adiposa, Endoconidiophora laricicola, E. polonica, and Penicillium freii DAOMC 242723 are presented in this genome announcement. These six genomes are from plant pathogens and otherwise economically important fungal species. The genome sizes range from 21 Mb in the case of Ceratocystiopsis minuta to 58 Mb for the basidiomycete Armillaria fuscipes. These genomes include the first reports of genomes for the genus Endoconidiophora. The availability of these genome data will provide opportunities to resolve longstanding questions regarding the taxonomy of species in these genera. In addition these genome sequences through comparative studies with closely related organisms will increase our understanding of how these pathogens cause disease.

Armillaria (Agaricales; Physalacriaceae) includes species of the most devastating fungal pathogens of trees and woody plants across the globe (Baumgartner ). Armillaria root rot, the disease caused by these species, manifests as shoot growth reduction, change in foliage characteristics, crown dieback and ultimately the death of natural forest and planted trees, as well as horticultural crops (Morrison ). Much attention has been afforded to the study of the taxonomy and biology of disease-causing Armillaria species. In an effort to understand the mechanisms of pathogenicity and infection of species that occur in the Northern hemisphere, the genome of A. mellea (Collins ) and transcriptome of A. solidipes (Ross-Davis ) have been sequenced. In addition, draft genomes of A. gallica and A. solidipes are available in the JGI MycoCosm (Grigoriev ) fungal genomes database. Despite the importance of Armillaria root rot in the Southern Hemisphere (Gregory ), the genome sequence of an Armillaria species native to that hemisphere has not yet been determined.

Armillaria root rot disease in the Southern hemisphere is caused by several species, including A. fuscipes (Gregory , Coetzee ). The species is restricted to the African continent where it affects the health of agronomic and timber plantations. In South Africa, Armillaria root rot disease was first reported during the early 1900s (Bottomley 1937), but A. fuscipes was only identified as the causal agent of root rot disease on Pinus species in 2000 (Coetzee ). Subsequent studies showed that this pathogen also affects the health of other tree and woody plant species elsewhere in Africa (Mohammed & Guillaumin 1993, Mwenje , Gezahgne ). Although research has been done to characterise African Armillaria species at the molecular level in terms of enzymes (e.g. Agustian , Mwenje & Ride 1996, 1997, 1999) - and genomic regions used for taxonomic purposes (e.g. Chillali , Pérez-Sierra , Coetzee ), nothing is yet known regarding the genome of A. fuscipes. The aim of this study was, therefore, to assemble a draft genome of this fungus with the objective of using the genome in future comparative genome studies.

SEQUENCED STRAIN

South Africa: Pinus elliottii, 2001, M.J. Wingfield (culture CMW 2740 = CBS 118115; PREM 61366 – dried culture).

NUCLEOTIDE SEQUENCE ACCESSION NUMBER

The Whole Genome Shotgun project of the Armillaria fuscipes genome has been deposited at DDBJ/EMBL/GenBank under the accession number LWUH00000000. The version described in this paper is version LWUH01000000.

METHODS

Isolate CMW 2740 was grown in liquid MYA (2 % Malt extract, 0.2% Yeast extract) at 24 oC in the dark for 6 wk. DNA was extracted from the harvested mycelium using a DNeasy Plant Mini Kit (Qiagen, Aarhus). Whole genome sequencing was performed on an Illumina HiSeq platform with a mate pair insert size of 250 bp, as well as PacBio sequencing with 20 single-molecule real time sequencing (SMRT) cells at the UC Davis Genome Centre (University of California, CA).

Quality scores, nucleotide composition, GC composition, and sequence duplication levels of the Illumina reads, were assessed with FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/), which resulted in the removal of the first 10 bp and last 15 bp of all reads. High quality short reads were employed to conduct error correction on the PacBio reads using the Celera Assembler (CA) pipeline (Miller ). For efficient error correction of the PacBio reads, coverage of 50x high quality short reads was used. For the A. fuscipes genome, 32 million high-quality Illumina reads of length 75 bp were used in the error correction. Assembly of the Illumina sequences was conducted using Velvet optimizer (Victorian Bioinformatics Consortium 2012) and CLC Genomics Workbench v. 5.5.1 (Qiagen). The error corrected PacBio data was assembled with VelvetOptimizer, CLC Genomics Workbench v. 5.5.1 and MIRA (http://www.chevreux.org/projects_mira.html). The quality of the assemblies was estimated based on the N50 value, number of contigs, and by applying the core eukaryotic gene mapping approach (CEGMA) (Parra ) as well as assessing single-copy orthologs with BUSCO (Simão ). Based on these metrics, the best assemblies from the corrected PacBio and Illumina read datasets were chosen to create an improved hybrid assembly using Graph Accordance Assembly program (GAA) (Yao ). Scaffolds were generated with SSpace (Boetzer ). Gene predication utilized RepeatMasker v. 0.10.1 (described in Tarailo-Graovac & Chen 2009) to remove repeat and low complexity DNA regions followed by Augustus (Keller ).

RESULTS AND DISCUSSION

The draft genome of Armillaria fuscipes (CMW 2740) was estimated at 53 Mb and the combined PacBio – Illumina assembly yielded 24 436 contigs with a N50 of 5 415 nucleotide bases. The GC content was 46.8 %. The CEGMA score indicated a completeness of 91.73 %, while 82 % (1190 of 1438) full-length single-copy orthologs were identified with BUSCO using gene models from Laccaria bicolor. In total, 199 (13 %) of these orthologs were fragmented and 49 (3.4 %) were missing. In silico gene prediction with Augustus yielded 14 515 genes with an average length of 1 350 bp and an average of 5 introns per gene.

Comparison of the A. fuscipes draft genome assembly showed that the genome size and the number of predicted genes are mostly congruent with other species of Armillaria (Table 1). The draft genome of A. gallica (JGI, MycoCosm, Grigoriev ) differed greatly from A. fuscipes, A. mellea (Collins ) and A. solidipes (= A. ostoyae) (JGI, MycoCosm, Grigoriev ) in terms of genome size and number of protein-coding genes (Table 1). The estimated genome size of A. fuscipes (53Mb) was somewhat smaller than the 58 Mb determined for A. mellea and A. solidipes. The number of protein-coding genes correlated with that of A. mellea, but was less than that determined for A. solidipes. Comparison of average gene length, exon length, intron length, and number of exons per gene, revealed similar results for all four Armillaria species.

Table 1.

Comparison of Armillaria fuscipes genome characteristics with those of other Armillaria species and members of the family Physalacriaceae.

	A. fuscipes	A. mellea (Collins et al. 2011)	A. solidipes (JGI Project Id: 1032462)	A. gallica (JGI Project Id: 1032466)	C. torrendii (Floudas et al. 2015)	F. velutipes (Park et al. 2014)
Genome Assembly (Mb)	53.00	58.30	58.01	85.34	31.57	35.6
Number of protein-coding genes	14 515	14 473	20 811	25 704	13 940	12 218
Average gene length (bp)	1 350	1 575	1 580	1 459	1 655	2 294
Average exon length (bp)	211.75	217.52	227	217	274	245.7
Average intron length (bp)	68.94	73.62	64	68	59	180.4
Average number of exons per gene	6	6.02	5.69	5.40	5.17	5.81

The genome assembly of A. fuscipes, as well as thos of the other Armillaria species, were compared to the publicly available assembled genomes of other members of the Physalacriaceae, namely Cylindrobasidium torrendii (Floudas ) and Flammulina velutipes (Park ) (Table 1). Cylindrobasidium torrendii and F. velutipes were characterised with smaller genomes than the Armillaria species, but the number of protein coding genes, average exon length, and average number of exons per gene ranges were comparable (Table 1). The genome of F. velutipes deviated from the other fungal genomes included in this study by having an average gene length that is much larger (2 294 in comparison to 1350–1655 bp) and an exceptional average intron length (180 in comparison to the 59–73.62 bp for the other species).

The genome sequence of A. fuscipes determined in this study is considered a high quality draft genome by virtue of the number of genes that were identified using the CEGMA pipeline and BUSCO analyses as well as the correlation in gene characteristics with other Armillaria species and members of the family Physalacriaceae. Armillaria fuscipes forms a basal taxon in the phylogeny of Armillaria species (Coetzee ) and its genome sequence therefore provides a crucial resource for future comparative genome studies.

Ceratocystiopsis minuta was first described as Ophiostoma minutum from Picea abies in Bialozieza, Poland (Siemaszko 1939). It is the type species of Ceratocystiopsis, one of the six currently recognized genera of Ophiostomatales (de Beer & Wingfield 2013). Ceratocystiopsis minuta has a wide distribution, with isolates reported from various countries on five continents (Siemaszko 1939, Davidson 1942, Mathiesen 1951, Mathiesen-Käärik 1960, Upadhyay 1981, Yamaoka , Zhou et , 2004a, 2004b, Kim , Plattner ). After carrying out a phylogenetic study on a collection of C. minuta isolates from various sources, Plattner suggested that isolates identified as C. minuta might represent a species complex of several phylogenetic species. The species boundaries and global distribution of C. minuta could thus be different from that as currently understood. Ceratocystiopsis minuta occurs in association with various bark beetle species that infest coniferous hosts (Siemaszko 1939, Mathiesen-Käärik 1960, Yamaoka , Zhou , 2004, Xudong et al. 2004, Kim ). An inoculation study with this fungus has shown that it is not a pathogen (Yamaoka ), and its relevance is likely to be simply as an agent of sap stain.

A number of genome sequences have been generated for species of Ophiostomatales. These include genomes from species of Leptographium s. lat. (DiGuistini , van der Nest , Wingfield ), Ophiostoma (Forgetta , Haridas , Khoshraftar ), Sporothrix (Cuomo , Teixeira , d’Alessandro ), and Graphilbum (Wingfield ). Of the current six recognized genera in Ophiostomatales, there was no genome sequence available for the genera Raffaelea or Ceratocystiopsis. The aim of this study was to generate the draft genome sequences of C. minuta, the type species of the genus Ceratocystiopsis and thus to provide a basis for comparison with other genera in the Ophiostomatales.

Austria: Ehrwald: isol. Ips cembrae, Aug. 1997, T. Kirisits (culture CMW4352 = CBS138717; PREM 61365 – dried culture).

The genomic sequence of Ceratocystiopsis minuta (CMW4352, CBS138717) has been deposited at DDBJ/EMBL/GenBank under the accession LZPB00000000. The version described in this paper is version LZPB01000000.

A single hyphal-tip culture was grown in 2 % malt extract and 0.5 % yeast extract broth at 25 °C, shaking at 150 rpm for several days. Mycelium was harvested and subjected to lyophilization. Genomic DNA was extracted from lyophilized mycelium using the method described by Aljanabi & Martinez (1997) with additional modification steps (Duong ). The methods of genome sequencing, assembly and annotation were similar to those used for Leptographium lundbergii (Wingfield ). Two pair-end libraries (350 bp and 530 bp average insert size) were prepared and sequenced using the Illumina HiSeq 2000 platform. Obtained reads were subjected to quality filtering and trimming and assembled using CLC Genomics Workbench v. 8.0.1 (CLCBio, Aarhus). The number of protein coding genes were predicted using the MAKER genome annotation pipeline (Cantarel ) following similar steps as for L. lundbergii (Wingfield ). The quality and completeness of the assembly were estimated using BUSCO (Simão ).

In total, over 23.1 million reads were obtained after quality filtering and trimming, with average read length of 85 bp. The assembled genome of C. minuta had an estimated size of 21.3 Mb, which was distributed in 904 scaffolds that were over 500 bp in size. The assembly had a N50 of 63.8 kb, with the longest scaffold was just over 398 kb in size. The mean GC content of the genome was 61.67 %, which is slightly higher than what has been reported for other species in the Ophiostomatales (DiGuistini , Haridas , Wingfield , b). The assessment of the completeness of the assembly by using BUSCO with a fungal dataset resulted in a completeness report of C:96%[D:7.6%],F:2.5%,M:1.2%,n:1348 (C:complete [D:duplicated], F:fragmented, M:missed, n:number of genes). MAKER predicted a total number of 6789 protein coding gene models, of which 5060 were multi-exonic genes. The mean exon length was 646 bp and the mean intron length was 114 bp. From the currently available data, C. minuta has the smallest genome and highest gene density for any member of the Ophiostomatales. The genome sequence of C. minuta presented in this study is the first genome sequence generated for Ceratocystiopsis and will serve as a valuable resource to study it’s biology and phylogenetic relationship with other members in the Ophiostomtales.

Genome sequences are available for numerous species of Ceratocystidaceae (de Beer ), providing a valuable resource for genomic studies on this important group of fungi. The first of these genomes to be made available was for Ceratocystis fimbriata s. str., the type species of Ceratocystis and the causal agent of rot of sweet potato and other root crops (Wilken ). This was followed by the genomes of two other pathogenic Ceratocystis species (C. albifundus and C. manginecans) as well as species of the related genus Huntiella (H. moniliformis and H. omanensis) (van der Nest , b). Subsequently, draft genome sequences for an additional six species of Ceratocystidaceae have been released (van der Nest , Wingfield , b). An additional three, those for C. adiposa, Endoconidiophora laricicola and E. polonica are provided in this announcement. This brings the total number of Ceratocystidaceae genomes in the public domain to 14, including representatives of the genera Ceratocystis, Endoconidiophora, Davidsoniella, Thielaviopsis and Huntiella (van der Nest ; Wingfield , b). Although this resource has been available for only a relatively short time, it has contributed significantly to a diverse array of studies including work on the mating system of these fungi (Wilken , Wilson ) as well as studies on ecological adaptation (van der Nest ).

The aim of this study was to produce a draft nuclear genome sequence for C. adiposa. Best known as the causal agent of root rot on sugarcane (Butler 1906), this species has also been reported from both Prunus (Paulin-Mahady ) and Pinus (Talbot 1956) species. Since its first description as Ceratostomella adiposum (Sartoris 1927), this species has also been treated as Ophiostoma adiposum and E. adiposa, and in 1981 C. adiposa was also suggested to be synonymous with C. major (de Beer ). This long and confused taxonomic history has yet to resolve the exact placement of C. adiposa within Ceratocystidaceae (de Beer ). Therefore, apart from adding to the Ceratocystidaceae genome resource, the availability of a genome sequence for this species could also provide useful tools to address its taxonomy.

Japan: isol. ex Saccharum officinarum, T. Miyake (culture CMW 2574 = CBS 136.34; PREM 61367 – dried culture).

The Whole Genome Shotgun project of the Ceratocystis adiposa genome has been deposited at DDBJ/EMBL/GenBank under the accession no. LXGU00000000. The version described in this paper is version LXGU01000000.

Ceratocystis adiposa isolate CMW 2574 was obtained from the culture collection (CMW) of the Forestry and Agricultural Biotechnology Institute of the University of Pretoria and grown on 2 % malt extract agar (20 g ME and 20 g Agar, Biolab, South Africa) supplemented with 100 μg/L thiamine for 2 wk. The plates were subjected to a previously described DNA isolation procedure (Roux ), before the purified nucleic acid was sequenced at the Genome Centre, University of California at Davis (CA) using a Genome Analyzer IIx platform (Illumina). For sequencing, paired-end libraries with approximately 350 and 600 base insert sizes were produced, and used to produce reads with an average length of 100 bases. Poor-quality reads and/or terminal nucleotides were discarded using the software package CLC Genomics Workbench v. 8.5.1 (Qiagen, Aarhus). A draft genome sequence was produced using the de novo assembly option in CLC Genomics Workbench under default options. The contigs produced were subjected to scaffolding using SSPACE and to fill the resultant gaps using GapFiller (Boetzer , Boetzer & Pirovano 2012). Putative open reading frames (ORFs) were predicted using the online version of the de novo prediction software AUGUSTUS based on the gene models for Fusarium graminearum (Stanke ), while genome completeness was assessed with the Benchmarking Universal Single-Copy Orthologs (BUSCO) software on all contigs greater than 500 bases in size (Simão ).

The draft nuclear genome of Ceratocystis adiposa has an estimated size of 28 447 711 bp as assessed through the summation of all contigs. The genome assembly has a N50 value of 114 645, an average coverage of 107x with a mean GC content of 47 %. Of the 951 contigs produced during the assembly process, 644 were 500 bp or larger and were retained in the final contig list. This draft assembly is predicted to contain 6 830 ORFs based on an AUGUSTUS analysis and has a density of 240 ORFs/Mb. The assembly also appeared to have a high degree of completeness with a BUSCO score of 97 %. The assembly contained 1 405 Complete Single-Copy BUSCOs, 93 Complete Duplicated BUSCOs, 26 Fragmented BUSCOs and 7 missing BUSCO orthologs out of the 1 438 BUSCO groups searched.

The overall genome statistics (i.e. genome size, ORF density and GC content) for the C. adiposa assembly is typical of what has been reported previously for Ceratocystidaceae (Wingfield ). That our assembly consists of a relatively small number of contigs, combined with the high completeness score, will be advantageous to any downstream analysis of the C. adiposa genome, especially in studies aimed at resolving its unclear evolutionary relationship with other Ceratocystidaceae (de Beer ). Apart from contributing to the ever-growing Ceratocystidaceae genome resource, the draft assembly presented here should be a useful resource for studying the biology C. adiposa.

The family Ceratocystidaceae (Microscales) includes numerous insect-vectored, wood-staining and pathogenic species of significant economic importance (Seifert , Marín & Wingfield, 2006, de Beer ). This diverse group of fungi occurs on both angiosperms and gymnosperms and was initially treated together in the genus Ceratocystis due to superficial similarities in their morphological features and mutualistic insect associations (Wingfield , de Beer ). The taxonomy of these fungi has recently been revised using a phylogenetic species concept, in addition to morphological, ecological and biological characters (de Beer ). Accordingly, the Ceratocystidaceae now incorporates several genera, including Ceratocystis s.str.

One of the newly recognised genera, Endoconidiophora, corresponds to what was previously known as the “C. coerulescens complex” (de Beer ). This genus includes both E. polonica and E. laricicola, two of the most aggressively pathogenic fungi associated with the bark beetles (Coleoptera: Scoltyinae) of Ips typographus and Ips cembrae that infest spruce (Picea abies) and larch (Larix decidua), respectively (Christiansen & Horntvedt 1983, Redfern ). As opposed to many other species of Ceratocystidaceae that have non-specific associations with fungivorous or sapfeeding insects, Endoconidiophora species have specific mutualistic relationships with their bark beetle vectors (Kile 1993, Wingfield ). In these associations, the fungi are thought to facilitate growth and reproduction of the insect, while the insect disperses the fungi (Paine ). However, the role of these fungi in tree death has been questioned (Six & Wingfield 2011).

The aim of this study was to sequence the genomes of E. laricicola and E. polonica to facilitate comparative studies with the genomes of other member species of Ceratocystidaceae already in the public domain (Wilken , van der Nest , b, Wingfield , b, Belbahri 2015). This study is also part of a larger objective to expand our knowledge of the biology and evolution of the interactions among fungi, bark beetles and their plant hosts.

Endoconidiophora laricicola: United Kingdom: Scotland: Larix sp., Nov. 1997, D. Refern (culture CMW 20928 = CBS 100207; PREM 61368 – dried culture).

Endoconidiophora polonica: Norway: Picea abies, Jan. 1990, H. Solheim (culture CMW 20930 = CBS 100205; PREM 61369 – dried culture).

The draft genome sequence of Endoconidiophora laricicola (CMW 20928) and E. polonica (CMW 20930) have been deposited in DDBJ/ENA/GenBank under the accession numbers LXGT00000000 and LXKZ0000000, respectively. The versions described in this paper are the first version of each genome, versions LXGT0100000 and LXKZ0100000.

Genomic DNA of E. laricicola isolate CMW 20928 and E. polonica CMW 20930 maintained in the culture collection (CMW) of the Forestry and Agricultural Biotechnology Institute (FABI), University of Pretoria, South Africa and the CBS-KNAW Fungal Biodiversity Centre (CBS), Utrecht, The Netherlands were sequenced with the Illumina HiSeq2000 platform at the UC Davis Genome Centre, University of California, Davis (CA). Two runs with 350-bp and 530-bp paired-end reads were performed following standard Illumina protocols to generate sequences with read lengths of approximately 100 bases. Reads were quality controlled and trimmed using CLC Genomics Workbench v. 7.5.1 (CLCBio, Aarhus). A draft genome assembly was performed using the de novo assembly option in the CLC Genomics Workbench under default options. SSPACE v. 2.0 (Boetzer ) was used to assemble contigs into scaffolds and gaps were filled using GapFiller v. 2.2.1 (Boetzer & Pirovano 2012). The Benchmarking Universal Single-Copy Orthologs tool, BUSCO (Software v. 1.1b1 of May 2015) (Simão ) was used to assess the completeness of the assembled genome. Lastly, the assembly was submitted to AUGUSTUS (Stanke ) in order to predict putative open reading frames (ORFs) using the gene models of Fusarium graminearum.

The draft genomes of Endoconidiophora laricicola and E. polonica had estimated sizes of 32 785 225 and 32 461 618 bases with an average coverage of 93x and 82x, respectively. The final E. laricicola assembly consisted of a total of 879 contigs larger than 500 bases, while the final E. polonica assembly consisted of a total of 914 contigs that were 500 bases or more in size. The GC content of both these two genomes was 45 % and the N50 for E. laricicola and E. polonica were 77 789 and 86 326 bases, respectively. The BUSCO analysis defined the genomes as 98 % complete, indicating that most of the core eukaryotic genes were present. From the E. laricicola analysis, we observed 1 415 complete single-copy BUSCOs, 104 complete duplicated BUSCOs, 19 fragmented BUSCOs and only 4 missing BUSCOs out of the possible 1 438 groups searched from the dataset for the fungal lineage. From the E. polonica analysis, we observed 1 415 complete single-copy BUSCOs, 97 complete duplicated BUSCOs, 10 fragmented BUSCOs and only 13 missing BUSCOs out of the possible 1 438 groups searched from the dataset for the fungal lineage.

Based on size, gene number and GC content, the E. laricicola (33.3 Mb, 6 897 ORFs, 45 % GC) and E. polonica (33.5 Mb, 6 792 ORFs, 45 % GC) draft genomes resembled those of other species in the Ceratocystidaceae. These include Davidsoniella virescens (33.7 Mb, 6953 ORFs, 44.5 % GC), C. fimbriata (29.4 Mb, 7266 ORFs, 48 % GC), C. eucalypticola (31.3 Mb, 7353 ORFs, 47.9 % GC), C. manginecans (31.7 Mb, 7494 ORFs, 47.9 % GC), C. platani (29.2 Mb, 5963 ORF, 48.2 % GC) and C. albifundus (27.2 Mb, 6967 ORFs, 48.6 % GC) that infect eudicots. This similarity also extended to Thielaviopsis musarum (28.5 Mb, 6963 ORF, 49.2 % GC) and T. punctuala (28.1Mb, 5480 ORFs, 48.3 % GC) that infect monocots, as well as the non-pathogens H. moniliformis (25 Mb, 6832 ORF, 48 % GC) and H. omanensis (31.5 Mb, 8359 ORFs eudicots, 47.6 % GC) (Wilken , van der Nest , b, c, Wingfield , b, Belbahri 2015).

Availability of the E. laricicola and E. polonica draft genomes provides an additional genome resource for members of the Ceratocystidaceae. These genomes can be incorporated in future comparative genomics studies to investigate the biology and evolution of species in the Ceratocystidaceae, as well as the evolution of the family within the broader Ascomycota. This is especially in regard to the interactions of these fungi with their insect vectors and plant hosts.

Penicillium freii is classified in Penicillium sect. Fasiculata (Houbraken & Samson 2011) together with such well-known species as P. camemberti and P. caseifulvum (used for making camembert cheese), P. crustosum (cause of apple rot; penitrem A and roquefortine C producer), P. expansum (cause of apple rot), P. roquefortii (used for making blue cheeses; roquefortine C producer), P. solitum (cause of fruit and nut spoilage; compactin producer), and P. verrucosum (cause of grain spoilage; ochratoxin A producer) (Frisvad & Samson 2004), among many others. Penicillium freii is a grain spoilage organism commonly found on barley and wheat in colder climates such as Scandinavia, the UK, and Canada, and produces xanthomegnins (hepato- and nephrotoxins) and penicillic acid (antibiotic, antiviral, antitumor, cytotoxic and a possible mycotoxin) (Frisvad ). It is part of the P. aurantiogriseum species complex, a group of mostly grain species that produce a variety of mycotoxins and have slow growing, blue-green agar colonies, smooth to finely roughened conidiophore stipes and smooth walled, more or less globose conidia (Frisvad & Samson 2004). On Czapek yeast extract agar (CYA), it produces crustose colonies, blue-green conidia and abundant exudate droplets (Fig. 1). In this study, we sequenced and annotated a genome draft of P. freii DAOMC 242723.

Fig. 1.

Penicillium freii (DAOMC 242723). A. Colonies (from left to right), CYA, CYA reverse, MEA, MEA reverse. B. Colony texture on CYA. C. Colony texture on MEA. D–F. Conidiophores. G. Conidia. Bars: F = 10 μm (applies to D & E), G = 10 μm.

Canada: Ontario: Ridgetown, isolated from wheat flour, 15 July 2011, collected and isolated by A. Schaafsma, identified by C.M. Visagie & N. Yilmaz (culture DAOMC 242723 – dried specimen deposited at DAOM).

NUCLEOTIDE SEQUENCE AND RAW READS ACCESSION NUMBERS

This Whole Genome Shotgun project is deposited at DDBJ/EMBL/GenBank under accession LLXE00000000. The version described in this paper is version LLXE01000000. Raw reads were deposited in NCBI SRA (http://www.ncbi.nlm.nih.gov/sra) accession number SRR2146291.

Penicillium freii DAOMC 242723 was grown on Blakeslee’s malt extract agar for 7 d at 25 °C (Visagie ) after which a spore suspension was prepared by flooding the colony with 5 mL sterile dH2O. One mL spore suspension was used to inoculate 100 mL Blakeslee’s malt extract broth in shaking flasks incubated at 300 rpm at 25 °C for 6 d. Fungal material was collected by filtration and DNA extracted with the OmniPrep kit for fungi (G-Biosciences) following the manufacturer’s protocol. Whole-genome sequencing (paired-end with 101 base pairs, bp) was performed on an Illumina HiSeq 2500 with TrueSeq V3 chemistry at the National Research Council Canada in Saskatoon, Saskatchewan, Canada.

FastQC v. 0.10.1 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was used to check the quality of genomic reads. Using fastx_trimmer (part of the FASTX-Toolkit v. 0.0.13 (http://hannonlab.cshl.edu/fastx_toolkit/)), 10 bases from the 5’ end were trimmed to yield higher quality reads of 91 bp. Adaptor sequences were removed with Trimmomatic v. 0.33 (Bolger ). Prior to genome assembly, the optimal k parameter was calculated with KmerGenie v. 1.6950 (Chikhi & Medvedev 2014). Error correction was performed on the trimmed reads with BayesHammer (Nikolenko ). De novo genome assembly was performed with SPAdes v. 3.5.0 (Bankevich ) with the option to correct mismatches and short indels enabled. Scaffolds shorter than 1000 bp were discarded. Assembly statistics were generated with QUAST v. 2.3 (Gurevich ). The assembly was checked by alignment of the corrected reads onto the scaffolds using Bowtie2 v. 2.0.0 (Langmead & Salzberg 2012). Alignments produced by Bowtie2 in SAM format were converted to sorted BAM format by SAMtools v. 0.1.19 (Li ) and statistics for nucleotide coverage were generated with Qualimap v. 2.1 (Garcia-Alcalde ). To evaluate the completeness of our assembly, CEGMA v. 2.5 (Parra ) was run on the scaffolds to detect the percentage of conserved eukaryotic genes (CEG’s) and BUSCO v. 1.1b1 (http://busco.ezlab.org/) was run on the scaffolds using the fungal profile (Dec 19, 2014 release) to detect Universal Single-Copy Orthologs. Species identification was confirmed by microscopic observation and BLASTing the internal transcribed spacer (JN942696) and beta-tubulin (AY674290) barcode sequences of P. freii (Visagie ) against the assembled genomic scaffolds.

Genome annotation was carried out using webAugustus (Hoff & Stanke 2013) running Augustus v. 3.0.3 (Stanke ). Predicted proteins were compared against the UniProt/Swiss-Prot manually curated fungal protein data set using BLASTp v. 2.2.28+. The BLAST hits with e-values less than 1.0E-100 and similarity hits ≥ 90% were assumed to be orthologs and were given protein names in the annotation set. Annotations were validated using Genome Annotation Generator (http://genomeannotation.github.io/GAG/) and tbl2asn (http://www.ncbi.nlm.nih.gov/genbank/tbl2asn2/).

Illumina sequencing of 32 million reads of Penicillium freii DAOMC 242723 represented 3.2 Gbp of data, which were assembled into 1923 scaffolds. The whole assembly was 33.5 Mbp with a GC content of 47.4 %. The following statistics for the assembly were obtained: N50 was 61.8 Kbp; the longest scaffold was 325 Kbp; the median nucleotide coverage across the whole assembly was 78.5x. Assessment of the completeness of the genome using BUSCO groups for fungi resulted in values of C:98%[D:6.9%],F:1.1%,M:0.1%,n:1438 (C:complete [D:duplicated], F:fragmented, M:missed, n:genes) whereas scores of 95.6 % and 97.6 % were obtained from the complete and partial gene set respectively using CEGMA. After annotation and validation, 11 739 predicted gene models were obtained, with 11 418 complete (97.2 %) and 320 lacking a start codon, stop codon or both (2.7 %). Mean gene length was 1489 bp, mean exon length was 483 bp and mean intron length was 79 bp. This draft genome of P. freii represents a useful resource for comparative genomic studies and in particular should be useful for those interested in polyketide and other mycotoxin synthesis in Penicillium and related genera.