SEQADV = _struct_ref_seq_dif

For FG3.3 I decided to link only UNP entries in the SEQADV.
A complication: when the UNP accession code is too long to fit in SEQADV it is truncated
in SEQADV and does not work: 6byp, 6a02
In these cases, the full accession code is in DBREF1 + DBREF2 (always both).
I decided to handle those in "Sequences".
*** See dbref1.txt ***

From Rachel Kramer Green April 2021:

Here is the complete list of possibilities in the _struct_ref_seq_dif.pdbx_seq_db_name category:

"EMBL"
"GenBank"
"NDB"
"NORINE" - 3odL
 "PDB"
"PIR"
 "PRF"
 "RefSeq"
 "UniProt"

----------------
3 of these can be searched for at RCSB advanced search:

ID(s) and Keywords
  Accession Code(s) - Uniprot
  Accession Code(s) - GenBank
  Accession Code(s) - NORINE // peptides

UniProt is in nearly all entries, and GenBank in ~1,500.
NOR is in 214 entries.


