Calcium-Binding Protein Protocols Calcium-Binding Protein Protocols

238 Weljie and Heringa
these databases). Similar functionality to SRS is found in other tools such as the
Entrez system.
3. In certain cases, it may not be useful to align the entire sequences such as with
sequences of differing size or containing repeats. For example, if one uses a global
alignment routine and has sequences that vary from less than 100 amino acids
to greater than 1000, the program will generally distort the shorter sequences in
attempting to create the global alignments. In this case, it might be advantageous
to use a local multiple alignment method, such as DIALIGN (47). A local pairwise
alignment tool such as BLAST can also be used to find local matches, and only
these portions of the sequence of interest can then be aligned using a global alignment
routine. This might be particularly appropriate in examining EF-hand or
C2-domain proteins from different subfamilies, as these sequences also contain a
variety of noncalcium-binding domains. The well-annotated SWISSPROT database
will give the specific residues for each entry belonging to a particular
domain, and this can be useful for extracting only the calcium-binding domains.
It should be noted that for reliable alignments, smaller sequences (approx 30 amino
acids) generally require slightly higher identities (>25 %) for reliable alignments,
as compared to longer sequences. Repeats are also difficult as there may be certain
sections of proteins that align better with incorrect topology (e.g., the C-terminal
end of one sequence with the N-terminus of another).
4. The amino substitution weights are normally given as a 20 × 20 matrix, containing
the weights for all possible amino acid exchanges. A multitude of such matrices
are available, of which the most widely used are the classic PAM250 (50), the
BLOSUM62 (27), and the Gonnet (51) matrix. The insertion/deletion penalties
are used to decrease the alignment score when gaps need to be made to optimally
match the two sequences. Normally, a pair of gap penalties is used, consisting of
an opening penalty used once for each gap and an extension penalty applied to
each incurring gap position. In practice, recommended values for the gap penalties
are available for most amino acid exchange matrices (such as 10 and 1 for the
PAM250 or 12 and 1 for the BLOSUM62 matrix, for gap opening and extension,
respectively). For divergent sequences, lowered settings for the gap penalties
could be attempted. Some alignment programs, such as CLUSTAL or MULTAL
(52) employ optimized combinations of exchange matrices with associated gap
penalties, which makes it less straightforward to vary gap penalties and/or try
other residue exchange matrices.
5. Although multiple alignments can confer a wealth of information, numerous
avenues of further analysis are open to the researcher in which an alignment plays
a key role. It should be mentioned, however, that each mode of analysis often
requires specific types of sequences to be included in the alignment together with
associated information. For example, homology modeling requires sequences
whose 3-D coordinates are known, and some types of phylogenetic trees require
outlier sequences to be included. The successful generation of alignments useful
in these applications may require the researcher to perform the alignment several
times in an iterative fashion before conditions are fully satisfied, albeit care in