Paul Medvedev 
In https://rdcu.be/dXA1e, @rayanchikhi, @yoanndufresne and I comment on the decoupling of ‘population’ pangenome graphs from ‘application-specific’ pangenome graphs (1/2)
"On the one hand, a population pangenome graph faithfully represents the complete variation of haplotypes present in some population of interest. On the other hand, an application-specific pangenome graph reflects only some of the population variability and is instead tailored toward maximizing the performance of application-specific downstream tools."