I'm trying to get a sense for whether or not read length has an impact on alignment success when mapping to a non-reference/congener genome. I'm conducting a large, multi-species RNAseq experiment and we have an annotated genome for one focal species in the study; the plan is to align reads for multiple congeners to this reference. In designing the experiment, I've come to the question of whether or not to go with PE50 reads (cost effective) or PE100/150. The increased cost of longer reads might be justified if they improve the likelihood of unique mapping, but I'm having a hard time finding papers that address this specific question.
Any input would be greatly appreciated.
Edit: changed a word