Singletons are challenging because you lack the comparative data that can really drive the decision as to which genes are in or out, which starts are conserved across a cluster, etc.
When you annotate a Singleton, rely on 4bp overlaps for start choices, your gene prediction algorithm outputs, and solid functional data.
And don't worry, as we find new cluster members, we can make changes to older annotations when we have more comparative data to work from.
Edited 02 Mar, 2018 15:59