You might try what is described in this post for Sanger generated sequences, I think this should at least get you started analyzing your data!
As @colinvwood pointed out, Oxford Nanopore sequences have a higher error rate than say Illumina generated short reads, but not only this, they tend to suffer from different types of errors than short read sequencing technologies (see this paper). The QIIME 2 project thus far has focused on amplicon-based analyses, which means that most of our efforts have been directed towards building high quality tools appropriate for short read technologies. With the coming release of a shotgun metagenomics toolkit, we may need to turn more of our attention to tools for working with other sequencing data, but there is no active work being done in this direction yet.