How can you interpret the output of genetic distances in MEGA using the Kimura 2-parameter?
Are these percentages, decimals that can be multiplied to give a percentage, the number of substitutions per site or something else entirely?
Hi Alice,
When you estimate the genetic distances among the sequences of your alignment MEGA build a matrix with the pairwise distances and if you want the distances in percentage (%) just multiply by 100 the obtained values. MEGA allow you to export the matrix in Excel format so that is easy.
In order to estimate the genetic distances you have to use a nucleotid substitution model. These models allow you to calculate the number of differences accumulated between two sequences since their last common ancestor. For phylogenetic analyses the selection of the best model can be done using programs like JmodelTest;
https://github.com/ddarriba/jmodeltest2
I don´t know what is the objective of your analyses...for example, for DNA barcoding it was proposed the use of Kimura-2-parameters , altough some authors suggest the use of p-distance...
Good luck!
Hi Alice,
When you estimate the genetic distances among the sequences of your alignment MEGA build a matrix with the pairwise distances and if you want the distances in percentage (%) just multiply by 100 the obtained values. MEGA allow you to export the matrix in Excel format so that is easy.
In order to estimate the genetic distances you have to use a nucleotid substitution model. These models allow you to calculate the number of differences accumulated between two sequences since their last common ancestor. For phylogenetic analyses the selection of the best model can be done using programs like JmodelTest;
https://github.com/ddarriba/jmodeltest2
I don´t know what is the objective of your analyses...for example, for DNA barcoding it was proposed the use of Kimura-2-parameters , altough some authors suggest the use of p-distance...
Good luck!
The percent identity between two sequences is often called the "Hamming distance". The Kimura two parameter model of evolution, and other models of evolution modify the percent identity to give a different weight for different mutations. In the Kimura two parameter model trasitions (A G and C T changes) are wieghted differently than transversions ( A T, A C, G C and G T) because transitions are observed to happen much more frequently than transversions. The models also attempt to correct for multiple hits per site, so as the sequences become more saturated with mutations there is a greater chance that an observed distance is less than the true distance between the two sequences. The Kimura distance between sequences is thus proportional to the % identity, but if you want simple pairwise identity just calculate that instead of trying to re-create it from the Kimura distance.
http://www.life.umd.edu/labs/delwiche/bsci348s/lec/NTSeqEvol.html
https://en.wikipedia.org/wiki/Models_of_DNA_evolution
- Badges
- Science topic
- Similar topics
- Biological Science
- Genetics