For region-based annotations, my understanding is that we can use either start+end pair or targets list to anchor annotations to the primary text data (as described in the vocab). However, current lif schema does not have targets field properly defined.