Biology Reference
In-Depth Information
representatives. These constraints are subject to consistency
measure to derive consistency-based scoring function.
6. The
N
2
representatives are then progressively aligned by the
consistency-based scoring function, with the aligning order
following a UPGMA tree estimated for the representative
sequences.
7. The pre-aligned groups are merged to the MSA of the
N
2
representatives to form an MSA of
N
1
target sequences. Finally,
the clusters with highly similar sequences obtained at the initial
clustering step are merged to the MSA of
N
1
target sequences
to form the MSA of all input sequences.
3
PROMALS3D Usage and Practical Issues
1. PROMALS3D is available as a Web server as well as a down-
loadable package at
http://prodata.swmed.edu/PROMALS3D
.
2. PROMALS3D Web server allows input of both sequences and
structures. The Web server extracts sequences from input struc-
tures and combines themwith input sequences to form the final
input sequence set. The Web server also prepares structural
alignments for input structures and feeds them as structural
constraints to the PROMALS3D program. On the other hand,
the PROMALS3D downloadable package currently only takes
sequences as input.
3. If only structures are input to the PROMALS3D Web server,
the final alignment is a consistency-based multiple structure
alignment that integrates both structural
information and
homolog-derived sequence information.
4. Input sequences should be in FASTA format and should not
have identical names. Certain characters in sequence names are
changed to “_”, including space, tab, and *?''";&\|/{})(][$,
but “.” (dot) and “
” are kept.
5. PROMALS3D Web server [
25
] offers various options of cus-
tomization of the final alignment output, such as displaying the
alignment with sequences colored by predicted secondary
structures and showing a consensus sequence and positional
conservation indices [
34
].
6. The UPGMA tree built for target sequences is reported. Since
it is based on a very crude measurement of evolutionary dis-
tances, it would not serve well for phylogenetic purposes.
7. The structure database is regularly updated in an automatic
fashion. The structure database contains a nonredundant set
of structures from the PDB database. The CD-HIT program is
used to cluster sequences with 3D structures at the 70 %
