Aims of this new version of BIBI

Prokaryotes Sequence Based Identification

General

speeding up the process
decision-making oriented functionalities
easy maintenance
minimalist

Databases

simplification of the database construction
simplification of the database maintenance
automation of the database construction
improvement of the detection of Type Strains [Type Strain] in the database
additional information useful to the biologist decision (biovars, serovars etc. and hosts, symbionts)
monthly buiding up of all the databases
nomenclature stringent/lax databases
including information on phylogeny-nomenclature dyscongruences (this is only partially done)

Computing

Python language
use of previously written libraries and classes
use of linux most basic tools whenever possible
Python CGI scripts only
pure html site

Selection of candidates sequences

BLAST is use, as previously in BIBI
extraction of the desired number of candidate sequences from the database (Bi Bi was using a parsing of blast results) in a very short time

Alignment

use of Muscle [1]
correction of the "trailing nucleotides" problem

Tree building

NJ tree build as in Bi Bi (Clustalw)
minimum loss of information (value of gaps)

Tree drawing

no use of applets
pure html tree enabling easy links building and evolutivity
direct translation into pdf
tree is the main information medium, not only a drawing : enriched tree concept

Tree analysis

basic functions, a visual inspection of the tree is better
commentaries on the results to focuse on potential interpretation pitfalls
quality control
use of previously written python classes

Alignment viewer

no applet
pure html
Sequence colourist (Stéphane Vellay and JP Flandrois 2003)

Extra functions

Crude BLAST result
incorporation of BLAST results to the tree
temporary files access for further works
plenty of links to improve the decision making of the biologist

___

Sun Aug 17 11:33:56 2008