To do this, you need to explicitly store it, best is probably in top-level metadata. So, for instance:
treeSeqOutput("out.trees", metadata=Dictionary("tags", p1.individuals.tagL0, "ids", p1.individuals.pedigreeID))
Then, after loading the tree sequence, you'd extract this like:
tags = ts.metadata['SLiM']['user_metadata']['tags']
ids = ts.metadata['SLiM']['user_metadata']['ids']
... and then you could use this to figure out which nodes to simplify to. Here I'm outputting IDs, so that you can make sure you know which tags match which individuals in the tree sequence; and then remember that simplify wants nodes, not individuals.
--peter