sub.sample and normalize.shared

Theory behind mothur
1

Hi,

I have question about sub.sample and normalize.shared comands.

My data are composed of six different groups and in each group is different sequence number (no artifacts in sequence data). So if i want to analyse alpha and beta diversity i probably have to normalize my data.
Which comand is more appropriate when comparing different groups with different number of sequences (they both work with .shared file type)?

the other thing i am curious about is from Shloss SOP.
In shloss SOP the comand summary single contains parameter subsample

˝summary.single(shared=final.an.shared, calc=…, subsample=…)˝
is this the same as sub.sample comand on .shared file and than performing a summary single on file subsample.shared?

i am a little confused about this…

Thanks in advance,


Maja
2

Which comand is more appropriate when comparing different groups with different number of sequences (they both work with .shared file type)?

We prefer sub.sample because of rounding issues associated with normalize.shared

˝summary.single(shared=final.an.shared, calc=…, subsample=…)˝
is this the same as sub.sample comand on .shared file and than performing a summary single on file subsample.shared?

This command (and also summary.shared/dist.shared/etc), subsamples to the size you specify, does the calculations and then repeats 1000 times (iter parameter). After 1000 iterations, it calculates the average and the standard deviation of the 1000 subsamplings. This is the way to go for all alpha and beta diversity analyses.

Pat