Revisiting pangenome openness with k-mers

Pangenomics is the study of related genomes collectively, usually from the same species or closely related taxa. Originally, pangenomes were defined for bacterial species. After the concept was extended to eukaryotic genomes, two definitions of pangenome evolved in parallel: the gene-based approach, which defines the pangenome as the union of all genes, and the sequence-based approach, which defines the pangenome as the set of all nonredundant genomic sequences. Estimating the total size of the pangenome for a given species has been subject of study since the very first mention of pangenomes. Traditionally, this is performed by predicting the ratio at which new genes are discovered, referred to as the openness of the species. Here, we abstract each genome as a set of items, which is entirely agnostic of the two approaches (gene-based, sequence-based). Genes are a viable option for items, but also other possibilities are feasible, e.g., genome sequence substrings of fixed length k (k-mers). In the present study, we investigate the use of k-mers to estimate the openness as an alternative to genes, and compare the results. An efficient implementation is also provided.