I am wondering what the best practices are analyzing host-parasite mixed transcriptomes. If I have mRNA-Seq data from infected host tissues that contain both host and parasite genomes, which would be the best way to tease them apart? Mapping to a combined reference (host + parasite), or mapping to the host genome first, collecting unmapped reads and mapping to parasite genome? What would be the implications of these options? When I get the real data, I could run tests with both approaches, but maybe someone already have some thoughts on this here?