Figure captions
Fig. 1. Phenotypic diversity of Psilocybe subaeruginosa from
populations in Australia, including pilei that are conical, papillate,
or sinuate, and fruiting from diverse substrates, including grass, leaf
litter, moss, and wood. A. Ellendale, Tasmania. B. Dover, Tasmania. C.
Tasmania (image courtesy of Karen Keats). D. Ellendale, Tasmania. E.
Victoria (image courtesy of Tannar Coolhaas). F. Tasmania (image
courtesy of Karen Keats). G. kunanyi, Tasmania. H. Ellendale, Tasmania.
I. Western Australia (image courtesy of Otto). J. Victoria (image
courtesy of Tannar Coolhaas). K. Victoria. L. Bunya Mountains,
Queensland.
Fig. 2. Analyses of Australian populations of Psilocybe
subaeruginosa (n=85), and P. azurescens (n=2) and P.
cyanescens (n=1) from the northern hemisphere. A. Provenance map of
Australian collections used in the study. B. 2-dimensional plot of
Discriminant Analysis of Principle Components (DAPC) based on 6,757 SNPs
with indels and sites under LD removed from the dataset, and individuals
coloured based on K=7 in the barplot of 2C. C. DAPC exploring K-values
2–8 based on the same dataset in 2B. Facets reflect geographic sampling
from local populations. D. SplitsTree network based on 1,555,848
LD-corrected SNPs, including indels, with all genomes treated as
haploids. Individuals are coloured by populations defined in DAPC
analysis at K=7. Sibling haplotypes sampled from the same pileus are
labelled with the same letter. Edge length reflects genetic difference
and reticulation may indicate recombination. Figures B and C produced by
R packages adegenet, vcfR, and ggplot2.
Fig. 3. Phylograms from maximum likelihood searches based on translated,
aligned, concatenated genes at mating-compatibility loci. Individuals
are coloured based on structured populations in Fig. 2C. ConcatenatedSTE3 genes at the pheromone/receptor (PR) locus, with compatible
(black) and incompatible (red) crosses among connected haplotypes. B.
Relationships among concatenated HD1 and HD2 genes at the
homeodomain (HD) locus. Alleles at mating compatibility loci are mostly
private to the sampled populations, which indicates high allelic
diversity in Australia.
Fig. 4. Measures of differentiation, nucleotide diversity and selection
across contigs containing mate compatibility genes among populations ofP. subaeruginosa . A–C homeodomain (HD) locus (826,289 base
pairs). D–F pheromone/receptor (PR) locus (122,582 base pairs). A.
FST plotted in 10,000 base pair windows of the HD locus
as a measure of differentiation among populations. Line colour indicates
which population has been removed from the comparison of all
populations. FST values approach 0 in admixed,
recombinant populations, and high FST values may
indicate divergence or a lack of recombination. FSTdecreases when the South Australian, Tasmanian, and Victorian population
is removed from the comparison, which indicates genetic differentiation
from populations east of the Great Dividing Range. B. Nucleotide
diversity (pi) plotted across 10,000 base pair windows of the HD locus
as a measure of diversity within defined populations. Diversity is low
within all populations (pi < 0.2). C. Tajima’s D index plotted
in 10,000 base pair windows across the HD locus as a measure of
selection across all populations. Positive values may be a signature of
balancing selection, in particular negative frequency dependent
selection, in which multiple alleles are maintained in populations and
no allele becomes dominant. D. FST plotted in 3,500 base
pair windows of the PR locus, with similar levels of differentiation
across the locus in all populations. E. Nucleotide diversity (pi)
plotted across 3,500 base pair windows of the PR locus, with divergence
among populations driven by diversity in the South Australian,
Tasmanian, and Victorian population. F. Tajima’s D index plotted in
3,500 base pair windows across the PR locus, with support for balancing
selection based on positive values comparable to the HD locus.
Fig. 5. Analyses of genetic diversity at the psilocybin locus. A.
SplitsTree network of aligned coding sequences of psiD, psiK,
psiM, psiT1, psiT2, psiR and two paralogs of psiH (psiH1and psiH2 ). Individuals are coloured based on structured
populations in Fig. 1A. Sibling isolates sampled from the same pileus
are linked by dashed lines. Siblings that have different alleles at
psilocybin loci indicate heterozygosity in the parental genotype.
Siblings with more than two genotypes (in populations from Bunya and
Shelley) may reflect recombination within the psilocybin locus. B.
FST plotted across individual genes as a measure of
differentiation among populations at the psilocybin locus.
FST is comparable among populations. C. Nucleotide
diversity (pi) plotted across individual genes as a measure of genetic
diversity at the psilocybin locus. The Khancoban population is the most
genetically diverse at the psilocybin locus relative to other
populations. D. Tajima’s D index plotted across individual genes as a
measure of selection at the psilocybin locus. Most genes have neutral to
positive values of Tajima’s D, which indicates balancing selection or
maintenance of diversity at the psilocybin locus. Measures of genetic
diversity were plotted in 50 base pair windows of all genes exceptpsiM , which used a 30-base pair window.
Fig. 6. Phylogram from a maximum likelihood search of an alignment of 26
ITS types of Psilocybe subaeruginosa and related taxa, showing a
sister relationship to P. cubensis . Sequences were obtained from
genomes in the present study, and from GenBank for sequences of P.
allenii, P. azurescens, P. cyanescens, P. makarorae , and P.
weraroa . Sequence abundance for each ITS type is shown in the adjacent
haplotype networks for P. subaeruginosa and P. cubensis ,
with dashes indicating the number of parsimony informative sites between
ITS types. The ITS region is intraspecifically variable in P.
subaeruginosa , which is paraphyletic with respect to other related
taxa.
Fig. S1. SplitsTree network based on 76,076 amino acids aligned from 194
single copy orthologs in Psilocybe subaeruginosa and relatives.
Branch length is informative for genetic distance between points,
reticulation is indicative of recombination, homoplasy, or incomplete
lineage sorting.
Fig. S2. Heatmap of sibling relationships based on pairwise comparisons
of the AJK statistic. Several of the relationships are known siblings,
and based on likelihood values of the AJK statistic, pairwise
relationships ≥0.91 are an indication that isolates are siblings.
Fig. S3. Evolution of psiH paralogs in the psilocybin locus inPsilocybe subaeruginosa .
A. Maximum likelihood tree of positional orthologs of the threepsiH paralogs extracted from cluster loci using exonerate. Bold
alleles are annotated as pseudogenes, and all sequences in thepsiH3 clade were pseudogenes. Isolate BRIP75275 has a predicted
functional paralog in the psiH3 position and a predicted
pseudogene in the psiH2 position. B. Clinker plot of synteny and
relatedness of genes in the psilocybin metabolic pathway. Darker
connections between plots indicate higher percent nucleotide identity.psiH2 is likely a duplicated copy of psiH1 , andpsiH3 is likely a duplication of psiH2 .
Fig. S4. Haplotype network of SNP diversity in the mitochondrial genome
based on 1,334 SNPs. Sibling populations shared the same mitochondrial
genotype (e.g., populations from Bunya, Ravensbourne, and Clifton Hill
and Geelong).