**Background:** The genus Chrysosporium comprises over 121 accepted species within the order Onygenales, many of which are opportunistic pathogens and producers of industrially relevant enzymes such as keratinases. Despite their biotechnological and clinical importance, only a few Chrysosporium genomes are available. Chrysosporium keratinophilum (ex-type strain CBS 104.62), originally isolated from soil in Papua New Guinea in 1957 and linked to the sexual morph Aphanoascus keratinophilus, had not been genomically characterized. This study aimed to sequence, assemble, and annotate its genome and identify secondary metabolite biosynthetic gene clusters.
**Methods:** Genomic DNA was extracted using a modified DNeasy Plant Mini Kit. Sequencing was performed on Illumina NovaSeq6000 (150 PE) and PacBio Sequel I platforms. Hybrid assemblies were generated using SPAdes v3.13.0 and MaSuRCA v4.0.5, with the best assembly polished using POLCA. Assembly quality was assessed with QUAST v5.1.0rc.1 and BUSCO v5.3.1. Taxonomic identity was verified via average nucleotide identity (ANI) analysis using Pyani v0.2.11 (ANIb BLAST+) against seven Onygenales genomes. Gene prediction and annotation employed Barrnap v0.9, tRNAscan-SE v2.0.9, BRAKER2 v2.1.6 (GeneMark-ET and AUGUSTUS), InterProScan v5.55-88.0, and BLASTKOALA v2.2. Carbohydrate-active enzymes (CAZymes) were identified using Run_dbCAN v3 with DIAMOND. Pathogenicity-related genes were detected via DIAMOND v2.0.15 against the PHI database (e-value 1×10⁻⁵, ≥80% identity, ≥100 amino acids, ≥60% query and subject coverage). Biosynthetic gene clusters (BGCs) were predicted using AntiSMASH v6.1.1 with default settings.
**Key Results:** The best assembly (MaSuRCA) yielded a polished genome of 25.4 Mbp across 25 contigs, with an N50 of 2.0 Mb, a GC content of 49.09%, and a BUSCO completeness score of 96.0%. A total of 166 tRNAs and 24 rRNAs were predicted. Gene annotation identified 34,824 coding sequences and 8,002 protein sequences. Functional annotation assigned functions to 76.6% (Pfam) and 62.7% (SUPERFAMILY) of predicted proteins. The most prevalent Pfam families included WD domain G-beta repeat, protein kinase domain, and reverse transcriptase. Notably, 36 LysM domain genes were identified—the highest number reported within Onygenales. CAZyme analysis revealed 421 genes across 148 families: 156 glycoside hydrolases (GHs), 151 glycosyltransferases (GTs), 57 carbohydrate-binding modules (CBMs), 36 auxiliary activities (AAs), 18 carbohydrate esterases (CEs), and 3 polysaccharide lyases (PLs). The PHI analysis identified 83 putative pathogenicity-related genes (1.06% of total), with 35 associated with reduced virulence. ANI analysis confirmed C. keratinophilum CBS 104.62 as belonging to the genus Aphanoascus (ANI 81.19% with Aphanoascus verrucosus). AntiSMASH classified 27 BGCs into nine types: six NRPS, six T1PKS, three terpene, one indole, two T3PKS, one lasso peptide, one NRPS-like, one beta-lactone, and six hybrid clusters. Four BGCs were matched to known compounds: okaramine B (85% similarity), UNII-YC2Q1O94PT/ACR toxin I (100%), clavaric acid (100%), and dimethyl coprogen (100%).
**Clinical Implications:** While C. keratinophilum is considered an opportunistic pathogen, the PHI analysis suggests weak pathogenic potential, with most homologous genes associated with reduced virulence. The genome encodes multiple peptidase families (S41, M16, M28, M36, M3, M48) and DPP IV linked to keratin degradation, supporting its ecological role as a keratinophilic saprophyte. The identification of BGCs with 100% similarity to clavaric acid (an antitumor Ras farnesyl transferase inhibitor) and dimethyl coprogen (a siderophore) indicates significant biotechnological potential. The presence of pectin-degrading CAZymes (PL3_2, PL1_7, GH28) absent in pathogenic Coccidioides species reflects its soil saprophyte niche. This genome provides a foundation for comparative genomics within Onygenales and for future studies on secondary metabolite production.