-auto-fix for xrefs, vocabularies, displayName, duplicate and dangling elements removal, etc. The Normalizer replaces URIs of canonical objects, such as EntityReference, ControlledVocabulary, BioSource, PublicationXref, with standard URIs (e.g., http://identifiers.org/uniprot/Q06609 - for a protein reference). Xrefs are consistently re-used, where possible, and being assigned with a generated URI. Original URIs of pathways, physical entities, interactions are preserved. Normalizer does not re-validate the modified model (one can validate again); errors will refer to original URIs. Not all problems can be fixed automatically, and some issues are consequence of others. Syntax errors, shared or completely missing unification xrefs usually cannot be fixed automatically.