The metric used to calculate distances (which affect the Distance and Tree views) is specified in the Distance section of the Style panel. To open the section, click on the expand bar entitled Distance, or choose View > Style > Distance.
To change the distance metric, use the *Metric*drop-down menu. Note that metric selection is disabled for projects without sequences.
- Uncorrected Pairwise Distance – (the default metric) – The number of identical bases (or residues) divided by the number of bases being compared, ignoring any positions with gaps). Uncorrected pairwise distance can be converted to %Identity using the formula: %ID = 100 * (1 – distance).
- Kimura – (protein sequences only) – The Kimura model (1983), which should not be used for very divergent sequences, uses the formula: D = – ln (1 – p – 0.2 p2), where p is the uncorrected pairwise distance. The Kimura distance approximates the “PAM distance” used in MegAlign.
- Scoredist – (protein sequences only) – This model, developed by Sonnhammer and Hollich (2005) computes the alignment score between two sequences using the BLOSUM62 scoring matrix. The units for the ScoreDist function are “percent accepted point mutations” (PAM). Scores are converted to distances and normalized by the average scores of the two sequences matched to them. This approach works even for very divergent sequences, providing that there are overlapping residues.
- Tamura-Nei (1993) – (nucleotide sequences and low degrees of divergence only) – An estimate of divergence using the TN93 model of the evolutionary process. This model attempts to separately account for different rates of transversion mutations (e.g., purine ↔ pyrimidine) and the two categories of transition mutations (i.e., purine only A ↔ G; pyrimidine only C ↔ T). The frequencies of each type of nucleotide do not need to be the same.
Need more help with this?