The fruit bodies of D. spathularia was collected in Luk Keng, Hong Kong on 20 June, 2022. The fungal isolate was transferred from the edge of fruit bodies to potato dextrose agar (BD Difco™) plates using a pair of sterilized forceps. The remaining collected fruit bodies were snap-frozen with liquid nitrogen and stored in -80°C refrigerator. Grown hyphae from >2-week-old was then transferred to new plates for purification for at least three rounds. The identity of isolate, termed “F14”, was validated with the DNA barcode of Translation elongation factor 1 alpha (TEF-1α) gene using primer pairs EF1-1018F and EF1-1620R (Stielow et al., 2015) (Supplementary Information 1).

~1.5 g of mycelia of D. spathularia isolate was collected from the upper layer of agar culture and was ground in a mortar with liquid nitrogen. High molecular weight (HMW) genomic DNA was isolated with a CTAB treatment, followed by using NucleoBond HMW DNA kit (Macherey Nagel Item No. 740160.20). Briefly, the ground tissue was transferred to 5 mL CTAB buffer (Doyle & Doyle, 1987) with an addition of 1% PVP for 1 h digestion at 55°C. After RNAse A treatment, 1.6 mL 3M potassium acetate was added to the lysate, followed by two rounds of chloroform:IAA (24:1) wash. The supernatant was added with H1 buffer from NucleoBond HMW DNA kit for a final volume of 6 mL and processed according to the manufacturer’s protocol. The DNA sample was eluted in 80 µL elution buffer (PacBio Ref. No. 101-633-500) and its quantity and quality was assessed with NanoDrop™ One/OneC Microvolume UV-Vis Spectrophotometer, Qubit® Fluorometer, and overnight pulse-field gel electrophoresis.

DNA shearing was first performed from 5 µg HMW DNA in 120 µL elution buffer using a g-tube (Covaris Part No. 520079) with 6 passes of centrifugation at 1,990 x g for 2 min. The sheared DNA sample was purified using SMRTbell® cleanup beads (PacBio Ref.

No. 102158-300), from which 2 µL of sample was taken for quality check through overnight pulse-field gel electrophoresis and Qubit® Fluorometer quantification. A SMRTbell library was then prepared by following the protocol of the SMRTbell® prep kit 3.0 (PacBio Ref. No. 102-141-700). Briefly, the sheared DNA was repaired and polished at both ends, followed by A-tailing and ligation of T-overhand SMRTbell adapters. A subsequent purification step was processed with SMRTbell® cleanup beads and 2 µL of sample was taken and subject to quality check as mentioned above. Nuclease treatment was then proceeded to remove non-SMRT bell structures. A final size-selection step using 35% AMPure PB beads was processed to eliminate short fragments.

A final library preparation was performed with The Sequel® II binding kit 3.2 (PacBio Ref. No. 102-194-100) before sequencing. The SMRTbell library was proceeded with annealing and binding with Sequel II® primer 3.2 and Sequel II® DNA polymerase 2.2, respectively. SMRTbell® cleanup beads were used for further cleanup the library, to which diluted Sequel II® DNA Internal Control Complex was added. The final library was loaded at an on-plate concentration of 90 pM with the diffusion loading mode. Sequencing was performed on the Pacific Biosciences SEQUEL IIe System for a run of 30-hour movies with 120 min pre-extension to output HiFi reads with one SMRT cell. Details of the resulting sequencing data are listed in Supplementary Information 2.

~0.5 g of stored fruit body was ground into powder with liquid nitrogen and used for the construction of an Omni-C library by following the plant tissue protocol of the Dovetail® Omni-C® Library Preparation Kit (Dovetail Cat. No. 21005). The ground tissue was transferred to 4 mL 1X PBS and was proceeded to crosslinking with formaldehyde and digestion with endonuclease DNase I. The quantity and fragment size of the lysate was assessed with Qubit® Fluorometer and TapeStation D5000 HS ScreenTape, respectively. The qualified lysate was polished at DNA ends and ligated with biotinylated bridge adaptors, followed by proximity ligation, crosslink reversal of DNA and purification with SPRIselect™ Beads (Beckman Coulter Product No. B23317). The end repair and adapter ligation were performed with the Dovetail™ Library Module for Illumina (Dovetail Cat. No. 21004). The library was then sheared with USER Enzyme Mix and purified with SPRIselect™ Beads. The DNA fragments were isolated in Streptavidin Beads, from which the library was amplified with Universal and Index PCR Primers from the Dovetail™ Primer Set for Illumina (Dovetail Cat. No. 25005). Size selection targeting fragment size between 350 bp and 1000 bp was performed with SPRIselect™ Beads. The quantity and fragment size of the library was assessed by Qubit® Fluorometer and TapeStation D5000 HS ScreenTape, respectively. The resulting library was sequenced on an Illumina HiSeq-PE150 platform. Details of the resulting sequencing data are listed in Supplementary Information 2.

~1 g of mycelia of D. spathularia isolate was ground in a mortar with liquid nitrogen. Total RNA was isolated from the ground tissue using the mirVana miRNA Isolation Kit (Ambion), following the manufacturer's instructions. The RNA sample was subjected to quality control with NanoDrop™ One/OneC Microvolume UV-Vis Spectrophotometer and 1% agarose gel electrophoresis. The qualified sample was sent to Novogene Co. Ltd (Hong Kong, China) for 150 bp paired-end sequencing. Details of the resulting sequencing data are listed in Supplementary Information 2.

A de novo genome assembly was conducted with Hifiasm (Cheng et al., 2021) , which was screened with BlobTools (v1.1.1) (Laetsch & Blaxter, 2017) by searching against the NT database using BLAST to identify and remove any possible contaminations (Supplementary Information 3). Haplotypic duplications were discarded using “purge_dups” according to the depth of HiFi reads (Guan et al., 2020) . The Omni-C data were used to scaffold the assembly using YaHS (Zhou et al., 2022).

Gene model prediction was performed using funannotate (Palmer & Stajich, 2020) . RNA sequencing data were first processed using Trimmomatic (v0.39) and kraken2 (v2.0.8 with kraken2 database k2_standard_20210517) to remove the low quality and contaminated reads. The processed reads were then aligned to the soft-masked repeat genome using Hisat2 to run the genome-guided Trinity (Grabherr et al., 2011) with parameters "--stranded RF --jaccard_clip", from which 44,384 transcripts were derived. Gene models were then predicted together with the protein evidence from Dacryopinax primogenitus (GCF_000292625.1; Floudas et al., 2012) using funannotate with the following parameters “--protein_evidence GCF_000292625.1_Dacryopinax_sp._DJM_731_SSP1_v1.0.proteins.faa --genemark_mode ET --optimize_augustus --busco_db dikarya --organism fungus -d --max_intronlen 3000". The Trinity transcript alignments were converted to GFF3 format and were input to PASA alignment in the Launch_PASA_pipeline.pl process to generate the PASA models trained by TransDecoder, followed by selection of the PASA gene models using the Kallisto TPM data. The PASA gene models were then used for training Augustus in the funannotate-predict step. The gene models from several prediction sources, with a total of 54,275 genes from Augustus (4967), HiQ (4624), CodingQuarry (11762), GlimmerHMM (10843), pasa (11217), snap (10862), were passed to Evidence Modeler to generate the gene model annotation files. UTRs were then captured in the funannotate-update step using PASA to generate the final genome annotation files.

Transposable element (TE) annotation was performed by following the Earl Grey TE annotation workflow pipeline (version 1.2, https://github.com/TobyBaril/EarlGrey) (Baril et al., 2022) .

A total of 9.34 Gb HiFi reads were generated from PacBio sequencing (Supplementary Information 2). After scaffolding with 3.46 Gb Omni-C data, the D. spathularia genome assembly has a size of 29.2 Mb, scaffold N50 of 1.925 Mb and 92.0% BUSCO score (Figure 1B; Table 1), and 19 out of 24 scaffolds are >100 kb in length (Figure and 1C; Table 2). The genome size is similar to Dacryopinax primogenitus (29.5 Mb) (Floudas et al., 2012) and GenomeScope estimated heterozygosity of 5.09% (Figure 1D; Table 3). Gene model prediction generated a total of 11,510 protein-coding genes with an average protein length of 451 bp and a BUSCO score of 91.9%.

Repeat content analysis showed that transposable elements (TEs) account for 1.62% of the D. spathularia genome (Figure 1E; Table 4). The major classified TE was LTR retransposons (0.95%) and DNA transposons (0.12%) (Table 4).