Lev Zhivotovsky wrote:
>> Dear colleagues,
>> Would you please let me know what's known about how to calculate distance
> between two amino acids based on biochemical properties of the molecules ?
>> Additional question then is how to compute distance between two amino acid
> sequencies of equal length that differ from each at several positions.
>> Is there any computer program to get the distance matrix for a set of amino
> acid sequences ?
>> Thank you in advance,
>> Lev Zhivotovsky
A place to look for this information would be an issue of "Methods in
Enzymology" called something like "Computer Methods for Biomolecular
Sequence Analysis" (around 1997 or 1998). This issue contains several
article that explain this topic.
In essence: The distance of amino acid types i and j can be gotten from
the 20x20-distance matrix M: d(i,j) = M(i,j). M can be defined via
biophysical similarities or from observed mutation rates in multiple
alignments blocks (PAM) or similar ideas.
If two sequences s1 and s2 both have the same length n (no gaps), you
can simply sum up the pairwise similarities: d(s1,s2) = sum(i=1 to n)
M(aa_type(s1[i]),aa_type(s2[i])).
I hope this wasn't over-simplistic,
cheers,
gerals
--
Gerald Loeffler
Email: Gerald.Loeffler at vienna.at
Smail: Apollo Imaging, Marchettigasse 7, A-1060 Vienna, Austria
Phone: +43 676 3289588 (+43 1 5952333 34)
Fax: +43 1 5952333 20
Keywords: Java, CORBA, OOA&D, Databases, Bioinformatics,
Computational Biology, Computational Biophysics
