Open petermr opened 9 years ago
org.xmlcml.ami2.plugins.phylotree.ScientificNameList
now reads and processes raw names.dmp
from NCBI. We will gradually add new rules to filter out false names,
Would prefer to keep it species level for the time being (proposal A) rather than genus level (proposal B) - but yes good to mention the possibility.
I think the middle field may be useful for removing some higher level categories (families) but I/we don't understand it well enough yet.
The value of the genus
is that we can check for garbles. I wasn't suggesting we did trees at genus level yet.
oh, ok
taxdump
field 2 (0-based) is a collection of annotations collected in taxdump/class.txt
. Probably not useful. Fields with leading uppercase seem entry-specific. The lowercase ones are more generic but not systematic:
<aardvarks>
<acorn worm>
<actinomycete>
<actinopterygian>
<agaric fungus>
<agrotis>
<all crocodiles>
<all species>
<allotype>
<alpha subdivision>
<amphibia>
<amphibius>
<amphipod>
<anamorph Cystobasidiales>
taxdump
field 4 (0-based) is a set of categories ("roles") of which we cureently omit all except scientific name
:
acronym
anamorph
authority
blast name
common name
equivalent name
genbank acronym
genbank anamorph
genbank common name
genbank synonym
in-part
includes
misnomer
misspelling
scientific name
synonym
teleomorph
type material
i.e. entry only selected if:
"scientific name".equals(field[4])
Note: moved summary files back to taxdump
folder.
Future editing. We can either filter names.dmp
per record (e.g.): :
/^\|1234/ d
will remove record for id 1234
or add generic filters. Since we may wish to revise filters we should probably build a series of filters to be read and run every time the files are rebuilt. These will be regexes read by ScientificNameList
.
The NCBI taxdump contains ca 600K records, looking like:
I have looked through this in some detail and propose that for current
ami-phylo
(mainly tackling IJSEM) we use ONLY "scientific name"s. These name have the following forms:The qualifiers include:
and many more.
[a-z]+
(results ingenus
only)These are then sorted and duplicates removed, leaving only single-word genus or two word binomial.