2.4 Gene annotation and phylogenetic analysis
To increase the accuracy of gene prediction，GeneMark-ES v4.32 and AUGUSTUS v2.7 were used to obtain abinitio prediction of gene structure(Stanke and Morgenstern，2005；Ter-Hovhannisyan et al.，2008).The data were combined by Maker v2.31.8 to produce a consensus gene set(Holt and Yandell，2011).For functional annotation，protein sequences were blasted against a series of protein databases，including NCBI RefSeq(National Center of Biotechnology Information Reference Sequence)，UniProt KB(UniProt Knowledgebase)，Swiss-Prot，KEGG，KOG(Eu Karyotic Orthologous Groups)，GO，and Pfam.PKSs were predicted by SMURF(Secondary Metabolite Unique Regions Finder，http：//www.jcvi.org/smurf)and antiSMASH(antibiotics& Secondary Metabolite Analysis Shell，https：//fungismash.secondarymetabolites.org/).Phylogenetic trees for PKSand CAO were constructed using the maximum-likelihood approach implemented in MEGA v6.06(Tamura et al.，2013).For LDC，CAO，and BBE gene prediction，we downloaded all LDC，CAO，and BBE genes in NCBI and conducted whole-genome Blastp.The Fungal Cytochrome P450 database(http：//drnelson.uthsc.edu/cytochromeP450.html)was used to identify P450 by Blastp，with an e-value cutoff of 1e-50.All proteins containing one 2OG-FeII_Oxy motif(Pfam ID：PF03171)were regarded as 2OGD(Kawai et al.，2014).Orthologous genes between hy4 and Cg01 were identified by OrthoMCL(Li et al.，2003).(https://www.daowen.com)