Wyślij emailem: Massively parallel read mapping on GPUs with the q-group index and PEANUT