What Does It Mean That Blosum Does Not Model Real Substitutions (As Opposed To Say Pam?)

Question

[Click downvote]

A lecture slide included these three sentences, the third and last is cryptic to me:

BLOSUM:

Compares sequences directly (no phylogeny)

– Does not model real substitutions...

What does it mean?

e: I get that a difference is that PAM uses phylo trees, and BLOSUM uses multiple alignments, but why aren't the compared bases from BLOSUM "real" substitutions?

Answer 1

[Hamish]

At a guess this could be due to BLOSUM being based on observed substitution frequencies in conserved blocks from a wide range of proteins, and thus not being based on an evolutionary model. In contrast to PAM where there is a model of substitution frequencies based on whole protein homology, which allows extrapolation to greater evolutionary distances than those present in the source data. This cannot be done with BLOSUM.

The Wikipedia article "Substitution matrix" provides a reasonable overview of the differences and their consequences. For more detail the following may help:

• "A model of Evolutionary Change in Proteins" Dayhoff, M.O., Schwartz, R. and Orcutt, B.C. (1978). Atlas of protein sequence and structure
• "Amino acid substitution matrices from protein blocks." Henikoff, S. and Henikoff, J.G. (1992). PNAS 89 (22): 10915–10919. doi:10.1073/pnas.89.22.10915
• "BLOSUM62 miscalculations improve search performance" Styczynski, M.P., Jensen, K.L., Rigoutsos, I. and Stephanopoulos, G. (2008). Nat. Biotech. 26 (3): 274–275. doi:10.1038/nbt0308-274
• "Studies of Varying Alignment Algorithm, Amino Acid Scoring Matrix, and Gap Penalties" Mount, D.W. (2008). CSH Protocols; 2008; doi:10.1101/pdb.ip60