Comparative Genomics of Microbotryum Fungus

We compared M. lychnidis-dioicae to 18 other fungi (Additional file 13) that sample the three subphyla in Basidiomycota, including 5 other Pucciniomycotina, 7 Agaricomycotina, 3 Ustilaginomycotina, as well as 3 Ascomycota outgroups. For M. lychnidis-dioicae and the 18 other fungal genomes, we identified ortholog clusters using OrthoMCL [126 (link)] version 1.4 with a Markov inflation index of 1.5 and a maximum e-value of 1 × 10⁻⁵. Two genomes, R. glutinis and P. placenta, are missing more broadly conserved orthologs than the other genomes; examining the 961 Microbotryum gene clusters with an ortholog missing in just one other genome, the number of missing clusters in any one Basidiomycete genome ranged from 1 to 34 with the exception of R. glutinis and P. placenta, missing 410 and 393 of these highly conserved clusters, respectively. PFAM domains within each gene were identified using Hmmer3 [127 (link)], and gene ontology terms were assigned using BLAST2GO [128 (link)].
To examine gene duplication history, the phylome, or complete collection of phylogenetic trees for each gene in a genome, was reconstructed for Microbotryum lychnidis-dioicae and 19 other fungi, including those used for OrthoMCL (Additional file 13) and Serpula lacrymans. Phylomes were reconstructed using the previously described pipeline [129 (link)]. All trees and alignments have been deposited in PhylomeDB [129 (link)] and can be browsed on-line (www.phylomedb.org, phylome code 180). Trees were scanned to detect and date duplication events [130 ].
RNAi components from other other fungi were used as Blast queries to find homologs in M. lychnidis-dioicae; the queries used include U. hordei RdRp (CCF48827.1), C. neoformans Ago1 (XP_003194007), and N. crassa Dcl2 (Q75CC1.3) and Dcl1 (Q758J7.1). The putative function was confirmed by examining protein domains. The identified domains for each protein include: Piwi, PAZ and DUF1785 found in both copies of Argonaute (MVLG_06823, MVLG_06899); DEAD/DEAH helicase, double-stranded RNA binding, and RNAseIII (MVLG_01202). Sugar transporters were identified based on homology to the Ustillago maydis Srt1t transporter (Genbank: XP_758521) and the Uromyces viciae-fabae Hxt1 (Genbank: CAC41332).
The M. lychnidis-dioicae protein models corresponding to carbohydrate-active enzymes were assigned to families of glycoside hydrolases (GH), polysaccharide lyases (PL), carbohydrate esterases (CE), carbohydrate-binding modules (CBM), auxiliary activities (AA) and glycosyltransferases (GT) listed by the CAZy database [64 (link)], exactly as previously done for the analyses of dozens of fungal genomes [39 (link), 66 (link), 131 (link), 132 (link)].