Annotated genome assemblies for Arachnopeziza aurata and Arachnopeziza aurelia

Description

Methods

Whole genome sequencing and genome assembly

High molecular weight genomic DNA was obtained from A. aurata and A. aurelia, respectively, cultivated in PDA at 80 rpm and 28 °C for 7-14 days. Mycelia balls were flash-frozen in liquid N₂ and crushed to a fine powder using mortar and pestle. Then, DNA was isolated with a CTAB protocol according to (Feehan et al., 2017) with the modifications indicated in (Frantzeskakis et al., 2018). The DNA was further purified using the NucleoBond HMW DNA kit (Macherey-Nagel, Düren, Germany); DNA integrity was tested via a 0.6% agarose gel using ethidium bromide as intercalating dye and run at 30 V for 3 h. DNA quantity was determined using the Qubit dsDNA-BR assay kit on a Qubit 4 (Thermo Fisher Scientific, Langerwehe, Germany).

DNA shotgun sequencing was performed using Illumina NovaSeq (NovaSeq 6000) technology with 1 µg input DNA at the service provider CeGaT (CeGaT, Tübingen, Germany), yielding 150-bp paired-end reads. We trimmed raw reads using Trimmomatic v0.39 (Bolger et al., 2014) and assessed read quality with FastQC v0.12.1 (Babraham Bioinformatics, Cambridge, UK). Long-read sequencing was performed by MinION (Oxford Nanopore Technologies, Oxford, US) with R9.4.1 flow cells and the Ligation Sequencing Kit SQK-LSK112; basecalling was done using guppy v0.15.3. All raw reads are available at NCBI/ENA/DDBJ at project accession PRJNA1128938.

We generated draft genome assemblies using the long reads with Canu v2.2 (Koren et al., 2017), Flye v2.9.2 (Kolmogorov et al., 2019) with options ‘--iterations 3 --threads 12 --genome-size 42m --asm-coverage 50 -m 10000’, and NextDenovo v2.5.2 (Hu et al., 2024) with configuration options ‘sort_options = -m 10g -t 8’ and ‘nextgraph_options = -a 1 -q 10 -E 5000’, and then merged the assemblies with quickmerge v0.3 (Chakraborty et al., 2016) to obtain the best draft assembly. We then remapped the Illumina short reads to the respective merged assemblies using the function ‘bwa mem‘ of BWA v0.7.17-r1188 (Li & Durbin, 2009) and polished the assembly using pilon v1.24 (Walker et al., 2014).

We obtained basic assembly statistics using Quast v5.2.0 (Gurevich et al., 2013) and assembly quality estimations with CRAQ v1.0.9 (Li et al., 2023). Further, we identified the 5.8S, 18S, 28S nuclear ribosomal DNA (nrDNA) and ITS sequences using the nrDNA sequences of A. aurata CBS127674 and A. aurelia CBS127675 from GenBank (accessions MH864617.1, MH876055.1, MH864618.1, and MH876056.1). Genome completeness was estimated using 1,706 ascomycete core genes from the ascomycota_odb10 database with compleasm v0.2.6 (Simão et al., 2015; Huang & Li, 2023).

Telomere identification

Telomeres were manually identified at the ends of assembled contigs as telomeric repeats 5’-TTAGGG-3’ or 3’-CCCTAA-5’. To complete the sequence ends where telomeric repeats were not found, we used teloclip v0.0.4 (https://github.com/Adamtaranto/teloclip). Briefly, we mapped the long ONT reads to the respective assembly using Minimap2 v2.26-r1175 (Li, 2018) to retrieve nanopore reads at both ends of the assembly sequences containing telomeric repeats (5’-TTAGGG-3’) with options ‘-k 20 -ax map-ont’ and parsed the SAM files with SAMtools v1.18 (Li et al., 2009). Then, teloclip with options ‘--motifs TTAGGG,TTAAGGG --matchAny’ was used to filter reads mapping to chromosome ends, and with options ‘--extractReads --extractDir SplitOverhang’ to extract these reads. Then, the reads of each chromosome end were aligned via multiple sequence alignment using MAFFT v7.520 (Katoh & Standley, 2013) and Jalview v2.11.3.2 (Waterhouse et al., 2009) was employed to manually identify and extend scaffold ends via read alignment until the last aligning telomeric repeat where available.

Gene annotation

We used BRAKER3 v3.0.8 (Gabriel et al., 2021, 2023; Bruna et al., 2023) for evidence-based gene annotation of both A. aurata and A. aurelia. The respective RNA-seq data obtained in this work and the OrthoDB v11 Fungi protein dataset (https://www.orthodb.org/) (Kuznetsov et al., 2023) served as evidence datasets for BRAKER3 predictions. Reads were prepared for annotation by mapping to the respective genome assembly with HISAT2 (Kim et al., 2015) with ‘--max-intronlen 1000 -k 10’ and parsing the SAM files with SAMtools v1.18 (Li et al., 2009). We assessed the completeness of the gene annotations using BUSCO v5.5.0 (Simão et al., 2015) with ‘-m protein’ and the ascomycota_odb10 database.