Goal is to hand-edit NOTHING
Name of object is EM:acc_no
DNA sequence -> DNA object of same name
Database EMBL ID_name Acc_no
Upper region mapping:
---------------------
DE -> Title
DT -> Date DateType Text
Text is info on what done
(*) KW, split on ';' -> Keyword
Ian also splits DE and adds to Keyword. Then remove junk
words and look by hand (now).
(*) CC -> DB_remark
cDNA_EST from division on title line
Locus: we scan DE, KW, CC for worm gene name structure, then hand edit
For human can get this from GDB.
References:
-----------
use getz: EMBL -> Medline
Feature table:
--------------
Map the following to Subsequence objects, called {Name}.{feature}.1,
etc. all direct first generation children of the main object.
Following each are instructions on tags to be set in the subobject,
including mapping some qualifiers.
CDS if /pseudo then set Pseudogene
else set CDS
/codon_start=x -> CDS x
misc_RNA set Transcript
mRNA set mRNA
prim_transcript set Unprocessed_mRNA
rRNA set rRNA
/product=x -> rRNA x
scRNA set scRNA
/product=x -> scRNA x
snRNA set snRNA
/product=x -> snRNA x
tRNA set tRNA
/anticodon=(pos:x,aa:y) -> tRNA y
General qualifier mapping:
/gene -> Locus
(*) /note=x -> DB_remark "Note:x"
(*) /product -> DB_remark "Product:x"
(*) /gdb_xref=x -> Database GDB x
?Look at this - maybe just report to log file
in a distinct form another script can pull out.
Location mapping:
If position starts with '<', then set Start_not_found.
If position ends with '>', then set End_not_found.
Map join() to Source_exons.
All other Features get mapped to EMBL_feature, except:
Exon, Intron, if spanned by any Subsequence.
Source, but perhaps /Clone=x -> DB_remark "Clone:x",
and similarly for Tissue_type, etc. (*)
General rules for feature locations:
For EMBL_feature
(x..y) = a range, take the outside value, i.e. the one that
makes the feature longest.
U05316:1745..2396 really need to make a correctly ordered LINK object
and then propagate the Subsequence to be a subsequence of this
link. Very messy. Write message to log file, process by hand.