Fungi produce a remarkable diversity of chemical compounds to interact with their environment. Lichens—obligate mutualisms comprising fungi, photosynthetic organisms, and bacteria—produce a particularly wide range of secondary compounds with functions including UV protection, anti-herbivory, antimicrobial activity, and metal homeostasis. Despite this chemical diversity, the genetic basis and evolutionary processes underpinning most lichen secondary metabolites have not been investigated. In this thesis, I focus on photoprotective anthraquinone pigments in the diverse Teloschistales order (Ascomycota) as a case study to explore how adaptive metabolic traits arise and then diversify in lichen-forming fungi. Characterising the genes responsible for anthraquinone biosynthesis first required generating genome-scale data for the Teloschistales. Given the challenges of culturing lichen-forming fungi, I developed and implemented a lichen-specific metagenomics pipeline to sequence, assemble and annotate 24 new lichen-forming fungal genomes. Comparative genomics identified putative anthraquinone biosynthetic gene clusters (BGCs) in Teloschistales genomes and demonstrated that BGC diversification occurred via re-shuffling existing enzyme genes with novel accessory genes. To understand anthraquinone evolution across the whole clade, I then expanded my metagenomic approach to sequence all major Teloschistales lineages. I combined this genomic dataset with densely sampled multilocus data to produce a robust genome-scale time tree. Phylogenomic analysis showed around half of current Teloschistaceae genera are not supported, and I propose a set of stable, evolutionarily relevant higher taxa instead. To understand how genomic variation affects the metabolite phenotype, I jointly analysed the genomes with new untar- geted metabolome data. This revealed a complex interplay between genomic and metabolic variation and suggested that, for anthraquinones, BGC variation affects compound regulation and transport more than structural diversity. Finally, as anthraquinones are broadly cytotoxic, I hypothesised that anthraquinone-producing Teloschistaceae lichens evolved resistance mechanisms to avoid self-toxicity. Combining enzyme assays, axenic culture experiments, selection analysis and in silico protein modelling indicated that Teloschistaceae lichens achieved self-resistance through the evolution of efflux pumps, toxin methylation and resistant target enzymes. Together, my results demonstrate the power of multi-omic approaches to investigate the evolutionary processes that shape metabolite diversification in lichens.