I have a theoretical question with mothur! I want to know what kind of distance is used in mothur for clustering? I have been searching in the literature and found many measures, such as Kimura, Tamura, Gamma, etc. I was told in BioStars forum that mothur uses a specific distance, so how can I know more about this distance?
Thanks
Do you mean for generating OTUs? The sequence-sequence distance calculations are given in the dist.seqs page of the wiki. This isn’t really an area I know a lot about, so if the calculation scheme is actually one of the ones you mentioned then sorry.
Once you have a distance matrix I believe mothur uses UPGMA to cluster sequences into OTUs, based on this paper.
Pat may correct me though.
@dwaite has given you the appropriate document for the distance calculation, but I wanted to add a small comment…
I have a theoretical question with mothur! I want to know what kind of distance is used in mothur for clustering? I have been searching in the literature and found many measures, such as Kimura, Tamura, Gamma, etc. I was told in BioStars forum that mothur uses a specific distance, so how can I know more about this distance?
the measures you list were derived for protein coding sequences where gaps are considered missing data and are ignored. So, if you use kimura and one sequence has a G and another has a gap, that difference is ignored. This will significantly suppress the difference between the sequences, which is probably a bad thing. Also these corrections for multiple substitutions generally have their greatest effect at larger distances than we are interested in for OTUs (ie. 0.03). You can read more about all of this in our paper at: