Litter-decomposing Agaricales play key role in terrestrial carbon cycling, but little is known about their decomposition mechanisms. We assembled datasets of 42 gene families involved in plant-cell-wall decomposition from seven newly sequenced litter decomposers and 35 other Agaricomycotina members, mostly white-rot and brown-rot species. Using sequence similarity and phylogenetics, we split the families into phylogroups and compared their gene composition across nutritional strategies. Subsequently, we used Raman spectroscopy to examine the ability of litter decomposers, white-rot fungi, and brown-rot fungi to decompose crystalline cellulose. Both litter decomposers and white-rot fungi share the enzymatic cellulose decomposition, whereas brown-rot fungi possess a distinct mechanism that disrupts cellulose crystallinity. However, litter decomposers and white-rot fungi differ with respect to hemicellulose and lignin degradation phylogroups, suggesting adaptation of the former group to the litter environment. Litter decomposers show high phylogroup diversity, which is indicative of high functional versatility within the group, whereas a set of white-rot species shows adaptation to bulk-wood decomposition. In both groups, we detected species that have unique characteristics associated with hitherto unknown adaptations to diverse wood and litter substrates. Our results suggest that the terms white-rot fungi and litter decomposers mask a much larger functional diversity.

The four tables S5, S7, S9, S10 are part of the supplementary material containing data about gene copies across species and gene families (S5), statistical analyses results (S7), subclassification of gene families AA1, AA2, AA5, GH3 (S9), and the OrthoMCL results (S10). 

The original protein datasets (.txt files) were collected using Pfam, IPR, superfamily domain information, and blastp searches. Those files starting with A_ were used for phylogenetic analyses and OrthoMCL, while those file starting with B_ were only analyzed on the first principal component analysis, but not further for phylogenetics and OrthoMCL. 

After the initial alignments low quality models were manually removed. Those files have this format: FILENAME_excluded.txt

The datasets were alignned using MAFFT (FFT-NS-i) and manually examined using AliView in order to remove poorly aligned regions. The final aligned files are those in PHYLIP format (.phy).

The phylogenetic trees generated using RAxML at Cipres (PROTCAT and WAG) using 1000 bootstrap runs (.tre files)