I don’t know were to put my question so I start here.
For the interpretation of my data I want to compare a set of 16S clone libraries with some data sets published in 2008 by Santelli et al. (Abundance and diversity oy microbial liafe in ocean crust).
I downloaded the Satelli seqs from NCBI and processed them like described in their paper - with the difference that I’m using Mothur 1.29 and Santelli used DOTUR. But whatever I do, I can’t reproduce their rarefaction curves or richness estimators like ACE, Chao1 or Jacknife. I allways end up by a factor of 2 or 3 higher.
So my question is: Is it me how is doing something wrong or is there a difference between the programms Mothur und DOTUR?
It could be a thousand things - sequence curation, alignment method, many more. They say they aligned and edited their sequences in ARB - that could mean many things too.
DOTUR was really just the cluster command (using method=furthest as the default) plus the summary.single and rarefaction.single commands. mothur has much much more. Two differences in mothur is that mothur uses the average neighbor by default in cluster. You can change this to furthest using the method option in cluster. This change would make mothur output fewer OTUs than DOTUR. Also, mothur uses a hard cutoff on the distances where DOTUR rounded the cutoffs. So for mothur a cutoff of 0.03 means 0.03 and for DOTUR a cutoff of 0.03 could mean 0.0349. You can play with this by setting the hard option in cluster to F. It’s unlikely that this really matters much.
You can always email the authors and ask them for their data and see if they have the inputs.