Single-cell RNA-sequencing is a transformative new technology for detailed understanding of cancer biology. Tissue dissociation and transcriptional profiling of single cells can uncover the complex cellular makeup of individual tumours and cell specific expression patterns that can be lost by bulk gene-expression analysis. However, the logistics of processing and preserving human tissue samples using methods compatible with single cell analysis is not trivial. For rare cancer types these issues are further compounded by the infrequent opportunity to bank fresh tumour samples. Single-nuclei RNA-seq (sn-RNA-seq) is an alternative single cell analysis approach that enables the use of snap frozen tissue samples stored in freezer archives that have been banked over many years. We have optimised sn-RNA-seq on the 10x Chromium platform and applied this method to phaeochromocytoma and paraganglioma (collectively PPGL), a neuroendocrine tumour originating from sympathetic and parasympathetic tissues. PPGLs are remarkable for their hormone-producing functionality, high heritability and genetic diversity. By bulk gene-expression profiling, these tumours cluster into at least six distinct gene-expression subtypes that correspond to germline and somatic gene driver mutations and implicated cellular pathways. We have used sn-RNA-seq to analyse 30 snap-frozen PPGL tumours and two normal tissues of the adrenal medulla. These samples represented tumours of both sympathetic and parasympathetic tissues harbouring mutations in 13 different PPGL driver genes. We have identified a diversity of cell types in PPGLs as well as previously unrecognised genotype-specific gene-expression patterns, cell-cell signalling pathways and putative therapeutic targets. Single-nuclei RNA-seq is therefore a breakthrough method for the analysis of rare and heterogeneous tumours.