Detecting Positive Selection at Single Sites (Suzuki and Gojobori 1999)

---

Selective forces operating at the amino acid sequence level have been detected mainly by comparing the nonsynonymous substitutions per site with that of synonymous substitutions per site with that of synonymous substitutions per site. Generally speaking, the excess number of synonymous substitutions was considered to be the result of negative selection, whereas that of nonsynonymous substitutions was attributed to positive selection.

Suzuki and Gojobori (1999) proposed a method to detecting positive selection at single amino acid sites, assuming that the synonymous change was almost neutral. The method can not be used when your sequences are non-coding sequences

Most algorithms used in the package are identical with those in Suzuki and Gojobori (1999). The package does not implement the proposed rule, that is when the number of combinations for possible ancestral codons over all nodes exceeded 10,000, that site will be excluded. Moreover, we changed the genetic distance for Neighbor-Joining (NJ) tree, and "revised" Hartigan's method. The reason for the latter is presented here.

Consider a tree with four sequences or OTUs. The status of OTU are presented at Fig.1 . The possible ancestral status are presented, too. However, you can not get all combinations between two neighboring nodes because some of them do not fit the minimum mutation step. For example, if the ancestral status of Node a (Fig. 2) is 'T', then the ancestral status of Node b is 'T', then the mutation step must be more than 3. Three is the minimum mutation step. To resolve the problem, we just reconsider the THEOREM 3 (Hartigan 1973). Sets VV(a) = {T}, then it is done. Do not use VV(a) from THEOREM 2. The VV(a) means the possible ancestral status at node a, and the notation is used by Hartigan.

Here, we note

• st -- the average number of synonymous sites throughout the phylogenetic tree
• nt -- the average number of nonsynonymous sites throughout the phylogenetic tree
• sc -- the total number of synonymous mutations throughout the phylogenetic tree
• nc -- the total number of nonsynonymous mutations throughout the phylogenetic tree

To detect the positive selection, given the fact of that, mut = 4 (sc) + 1 (nc), the probability of that sc is equal to or more than 4 is calculated by binomial distribution. It will be labeled by red color and asterisk when the probability is less than 5%.

To detect the negative selection, given the fact of that, mut = 4 (sc) + 1 (nc), the probability of that nc is equal to or more than 1 is calculated by binomial distribution. It will be labeled by blue color and asterisk when the probability is less than 5%.