[Bio-srs] Question about amino acid substitution matrices

[volker.weinberger at novartis.com]

Dear Steffen,

in fact there are even different versions of BLOSUM62, both for 1/3 bit 
scaling and 1/2 bit scaling.  Gap penalties need to be adjusted to 
scaling.

Discussion on the impact of different scaling on optimal gap penalties can 
e.g. be found in:
- J.T. Reese, W. R. Pearson: Bioinformatics (2002), 18:1500-1507, 
http://www.people.virginia.edu/~wrp/papers/bioinfo_wrp02jr.pdf
- W. R. Pearson: Profein Science (1995), 4:1145-1160 , 
http://www.proteinscience.org/cgi/reprint/4/6/1145 .

Best regards,
Volker






"Steffen Fehrmann" <Kasymir at gmx.net>
Sent by: bio-srs-bounces at oat.bio.indiana.edu
23.08.2005 19:50

 
        To:     bionet-software-srs at moderators.isc.org
        cc:     (bcc: Volker Weinberger/PH/Novartis)
        Subject:        [Bio-srs] Question about amino acid substitution matrices
        Category: 



Hi,

I'm a student of biology and one of my main focuses is bioinformatics.
At the moment I'm preparing for my final exam.
I read about that there are different bit units in different amino acid
substitution matrices, e.g. the BLOSUM62 Matrix, which is scaled in 1/2
bit units or the PAM 250, in 1/3 bits (Bioinformatics, Mount, D.W.). I
wouldn't care but gap penalty scores depend on these scales.
Despite I read how these matrices are generated, I didn't find a
description about the differences between 1/2 and 1/3 scaled aa
substitution matrices. So, what does that mean and how do the matrices
discriminate?
I hope anyone could help me and thank you for trying,

Steffen

_______________________________________________
Bio-srs mailing list
Bio-srs at net.bio.net
http://www.bio.net/biomail/listinfo/bio-srs