Substitution matrix

Henikoff S and J G Henikoff 1992. Amino Acid Substitution Matrices from Protein Blocks. Proceedir Hie National Academy of Sciences LISA 89 10915-10919. 

S Hemkoff, JG Henikoff. Performance evaluation of ammo acid substitution matrices. Proteins 17 49-61, 1993. 

This procedure led to a substitution matrix for aligning disordered protein that was different from the commonly used substitution matrices, such as BLOSUM62 (Fig. 4). The matrix for disordered protein is generally better than order-based matrices for aligning disordered proteins whose sequence identities are between 20 and 50%. These results indicate that disordered and ordered protein can be distinguished by their patterns of evolutionary change. 

Substitution matrices have also been modeled on amino acid or protein structural properties rather than on alignment data (Grantham, 1974 Miyata et al., 1979 Rao, 1987). However, the relationship between scores and target frequencies means that any imaginable set of scores has an implicit set of target frequencies, and obtaining these from the alignment data itself is the most direct approach (Altschul, 1991). 

Henikoff, S., and Henikoff, J. G. (1992). Amino acid substitution matrices from protein blocks. Proc. Natl. Acad. Sci. U.S.A. 89, 10915-10919. 

When the defining set is small, there is an inadequate representation of the range of allowed positional variation in terms of either amino acid frequencies or types. Pseudo counts have been used for estimates of such unobserved data (Brown et al., 1993 Lawrence et al., 1993 Claverie, 1994 Tatusov et al., 1994 Henikoff and Henikoff, 1996 Sjo-lander et al., 1996). Such pseudo counts have been based on background frequencies, existing substitution matrices (Henikoff and Henikoff, 1992), and Dirichlet priors (Brown et al., 1993 Tatusov et al., 1994 Sjolander et al., 1996). Tatusov et al. (1994) reported that the Dirichlet priors (Brown et al., 1993 Sjolander et al., 1996) were most effective in... 

Baldi, P. (1995). Substitution matrices and hidden Markov models. /. Gomput. Biol. 2, 487-491. 

Altschul, S.F. (1991) Amino Acid Substitution Matrices from an Information Theoretic Perspective, J. Mol. Biol. 219, 555-565. 

Distant Evolutionary Relationships Can Be Detected Through the Use of Substitution Matrices... 

From the examination of appropriately aligned sequences, substitution matrices can be deduced. In these matrices, a large positive score corresponds to a substitution that occurs relatively frequently, whereas a large negative score corresponds to a substitution that occurs only rarely. The Blosum-62 substitution matrix illustrated in Figure 7 9 is an example. The highest scores in this substitution matrix indicate that amino acids such as cysteine (C) and tryptophan (W) tend to be conserved more than those such as serine (S) and alanine (A). Furthermore, structurally conservative... 

---