1

I am trying to perform an NMDS ordination on a large data set with well over 50 sample points across different time points and varying collection points. Each sample point has thousands of OTUs sequenced/assembled for it.

I understand that one way of adding ellipses is to manually add 'meta-data' which can be incorporated into the NMDS plot, as shown in this question.

However, I would like to add ellipses without having to go through the tedious and potentially error-prone process of manually adding meta-data as shown in the aforementioned question.

Is there any way that I can somehow 'automate' this process? My sample points are all titled in the same format, one typical example is '12.14.2011.NP'. 'NP' stands for one of the sample sites.

I would like to know how I can create ellipses for all of the 'NP' points- and all of the other different sample sites too. The are all named by a particular acronym.

Thanks.

Community
  • 1
  • 1
Frodo Baggings
  • 57
  • 3
  • 6

1 Answers1

1

If you have a variable with levels, names or other identifiers for ellipses, you can use those. If these identifiers are embedded in other strings, you should remove the unnecessary parts. See gsub, substring, strsplit to manipulate the strings. If all your names are of the type you gave, and you want to remove numbers and dots and leave characters, this will work:

gsub('[0-9.]', '', '12.14.2011.NP')

or for all names in data frame mydata:

gsub('[0-9.]', '', rownames(mydata))
Jari Oksanen
  • 3,287
  • 1
  • 11
  • 15